incorporation choices of ipo firms, the - nyu law … · the incorporation choices of ipo firms ......

THE INCORPORATION CHOICES OFIPO FIRMS

ROBERT DAiNS*

This Article presents the first evidence about the choice of corporate law and themarket for corporate charters at an initial public offering. Though firms are free toincorporate in any of the fifty states and are said to search for optimal legal rules,they appear to simply make a binary choice: Delaware or their home state. Feder-alism has thus resulted in a series of local markets with one national producer,rather than a nationwide "race to the top/bottom." This pattern raises questionsabout how firms choose a state of incorporation and suggests that there is a sub-stantial home-state advantage (or home bias). Professor Daines explores reasonsfor this home bias and reports evidence that lawyers play a key role in determiningstate of incorporation and that agency costs may affect the advice they provide.Professor Daines also examines other factors that affect a firm's domicile, includingvariation in state law and firm characteristics. Takeover laws do not appear to beimportant and there is some mixed evidence that state law may have networkqualities.

At the time of an initial public offering (IPO), entrepreneurs de-cide on the legal rights and governance protections they will offer in-vestors in the firm. The entrepreneur must decide on the cash flowand voting rights to offer investors, the extent to which the firm'smanagers will be exposed to (or protected from) the threat of hostiletakeover,' whether shareholders will have the right to sue managersfor breach of duty of care,2 whether to waive certain potential duty-of-

* Professor of Law, New York University School of Law. J.D., 1992, Yale Law School.A previous draft was entitled "How Firms Choose Corporate Law Rules" (1998). I

owe a larger than usual debt of gratitude to those who have helped me think through thesequestions: Barry Adler, Bill Allen, Yakov Amihud, Jennifer Arlen, Ian Ayres, StephenGillers, Henry Hansmann, Ehud Kamar, Mike Klausner, Lewis Kornhauser, Ricky Revesz,Roberta Romano and Alan Schwartz. Marcel Kahan in particular provided encourage-ment and guidance in this long overdue project. Dave Koster and Amir Tayrani providedexcellent research assistance. I benefited from presentations at Columbia, NYU, Toronto,Yale, Bill Allen's corporate law bridge group, and the American Law and Economics As-sociation, and from the assistance of the Filomen D'Agostino and Max E. Greenberg Re-search Fund. I am especially indebted to Ralph Winter and Roberta Romano, who firsttaught me (and interested me in) corporate law and state competition. My son Aaron, age9, has agreed to accept responsibility for all errors.

1 See Robert Daines & Michael Klausner, Do IPO Charters Maximize Firm Value?Antitakeover Protection in IPOs, 17 J.L. Econ. & Org. 83, 86-88 (2001) (examining use oftakeover defenses in IPO firms).

2 See, e.g., Del. Code Ann. tit. 8, § 102(b)(7) (2001) (permitting charter provisioneliminating or limiting directors' liability for breach of duty of care).

1559

Imaged with the Permission of N.Y.U. Law Review

NEW YORK UNIVERSITY LAW REVIEW

loyalty conflicts,3 how the firm's executives will be compensated, andhow the board of directors will be constituted. Typically, the entrepre-neur is assumed to design these legal rights and governance features ina way that maximizes the value of the firm.4 These rights and featuresare likely to be important because, on average, an IPO cuts managers'ownership by roughly 50%,5 and managers therefore will have imper-fect incentives to improve the value of the firm following the IPO.

One potentially important choice made at the IPO is the corpo-rate law that will govern investors' rights and managers' duties. In theUnited States, each of the fifty states, plus the District of Columbia,enacts its own corporate law and maintains its own court system forresolving corporate disputes. Firms can elect to be governed by any ofthese regimes simply by incorporating in the state of their choice,without regard to where they operate. 6 Firms are thus free to choosetheir corporate law. How do they choose?

In an earlier study, I found that firms incorporated in Delawarewere worth significantly more than firms incorporated elsewhere (per-haps by as much as 2% on average).7 Delaware firms also receivedsignificantly more takeover bids and were more likely to be sold.8

These findings are consistent with the theory that, relative to otherstates, Delaware better facilitates takeovers, thereby improving share-holder value. The findings received attention because they suggestthat Delaware law is not, on balance, harmful to shareholders-some-thing that has been long debated. 9 However, these findings are inter-

3 See, e.g., § 122(17) (permitting corporations "to renounce.., any interest or expec-tancy of the corporation in ... specified business opportunities ... that are presented to thecorporation").

4 See, e.g., Frank H. Easterbrook & Daniel R. Fischel, The Economic Structure ofCorporate Law 4-8 (1991) (explaining managerial efforts to create beneficial legal struc-tures); Steven Kaplan & Per Stromberg, Financial Contracting Theory Meets the RealWorld: An Empirical Analysis of Venture Capital Contracts 2 (Ctr. for Res. in Sec. Prices,Univ. of Chi. Graduate Sch. of Bus., Working Paper No. 513, 2000), http://papers.ssrn.com/so13/delivery.cfm/000320302.pdf?abstractid=218175 (discussing this assumption in venturecapital context).

5 See Daines & Klausner, supra note 1, at 93 tbl.1.6 Under the long-standing "internal affairs" rule, investors' rights and managers' obli-

gations are determined by the corporate law rules of the incorporating state and not therules of the state in which the firm is headquartered.

7 The sample included data on 4481 exchange-traded firms, including over 47,000 firmyears between 1981 and 1996. Robert Daines, Does Delaware Law Improve Firm Value?,62 J. Fin. Econ. 525, 532 (2001).

8 Id. at 541.9 See Roberta Romano, The State-Competition Debate in Corporate Law, 8 Cardozo

L. Rev. 709, 710-17, 720-25 (1987) (summarizing classic positions on state competition andexplaining why Delaware is destination state of choice); see also Easterbrook & Fischel,supra note 4, at 213-15 ("As a matter of theory, the 'race for the bottom' cannot exist.");Lucian A. Bebchuk, Federalism and the Corporation: The Desirable Limits on State Com-


[Vol. 77:1559

December 2002] INCORPORATION CHOICES OF IPO FIRMS

esting for a second reason-they suggest that incorporation choicesmatter and that there is something unique about Delaware incorpora-tion during this sample period. Delaware law either helps make firmsmore efficient or it attracts fundamentally different firms.10 If incor-poration state matters, the natural question is this: How do firms se-lect a state of incorporation? My prior paper ended with this questionand this Article is a first step towards an answer."

One might expect that the question of where firms incorporateand how they choose would be well studied, especially given the vigor-ous debate over whether firms should be allowed to choose where toincorporate or whether mandatory federal law should govern.12 Sur-prisingly, it is not. Legal scholars have examined the theoretical con-sequences of firms' choice of corporate law and have consideredgenerally how firms might select a legal regime, but they largely haveignored what firms actually do.' 3 Similarly, there has been little dis-

petition in Corporate Law, 105 Harv. L. Rev. 1435, 1441 (1992) (advocating greater federalregulation in various areas of corporate law); Lucian A. Bebchuk & Allen Ferrell, Federal-ism and Corporate Law: The Race to Protect Managers from Takeovers, 99 Colum. L.Rev. 1168, 1172, 1190-91 (1999) (arguing that Delaware rules governing takeover defensetactics are "characterized by unnecessary ambiguity and unpredictability"); Ronald J. Dan-iels, Should Provinces Compete? The Case for a Competitive Corporate Law Market, 36McGill L.J. 130, 138-40 (1991).

10 Many had considered corporate law trivial. See, e.g., Bernard S. Black, Is Corporate

Law Trivial?: A Political and Economic Analysis, 84 Nw. U. L. Rev 542,544 (1990) (argu-ing that "mandatory" corporate laws are actually trivial because they "do not prevent com-panies ... from establishing any set of governance rules they want"). However, differencesin valuation and takeovers in Delaware firms were durable and significant, even thoughfirms incorporated elsewhere might have adopted (by contract) many provisions of Dela-ware law. This suggests that domicile matters in ways contract cannot replicate.

11 See Daines, supra note 7, at 555-56. An earlier draft of this paper in 1998 asked howfirms choose a state of incorporation. After presenting the initial draft, however, it wasclear that many believed corporate law was uniform and the choice of domicile thereforetrivial, so I put the draft aside to finish the paper on Delaware firms. Now that it hasbecome clear that Delaware incorporation matters, how firms decide where to incorporateis a more interesting question and I take it up again here.

12 More generally, the debate is over whether firms should be allowed to choose their

own corporate law rules. See generally, Symposium, Contractual Freedom in CorporateLaw, 89 Colum. L. Rev. 1395 (1989) (addressing mandatory role of corporate law); see alsosupra note 9 (surveying various positions in debate over whether state competition harmsshareholders).

13 The exception is Roberta Romano, Law as a Product: Some Pieces of the Incorpora-tion Puzzle, 1 J.L. Econ. & Org. 225 (1985). Her study, however, focuses only on thereincorporation decision and not the original incorporation choice. See id. at 226. Financescholars have shown a greater interest in the actual governance choices of initial publicoffering (IPO) firms and how these choices may affect shareholders. Studies have ex-amined their adoption of takeover defenses, see Daines & Klausner, supra note 1, at 95-110 (examing potential efficiency justifications for takeover defenses in IPO firms); LauraCasares Field & Jonathan M. Karpoff, Takeover Defenses at IPO Firms, 57 J. Fin. (forth-coming 2002), http://papers.ssrn.comlsol3/papers.cfm?abstract id=286923, and the compo-sition of IPO boards, see Malcolm Baker & Paul A. Gompers, The Determinants of Board



cussion and no evidence about what states other than Delaware aredoing and what qualities make a state's corporate law attractive. 14

This Article is an initial attempt to fill this gap and answer ques-tions posed by my earlier findings about Delaware firms. My primarygoal is to describe the pattern of firm choice among IPO firms and toexamine how firms choose among alternatives. Other papers will takeup additional questions raised by these findings.

In some ways, my findings alter or are unexplained by popularconceptions of state competition. My central finding is that there islittle evidence of a nationwide market in legal rules. In spite of all thedebate about firms' freedom to incorporate anywhere, the importanceof corporate law, and spirited state competition for charters, firms'actual choices are much more mundane: 97% of public firms incorpo-rate either in their home state or Delaware.15 Firm choices are thusoddly "bimodal"-they operate as if there is no national market but asingle choice: their home jurisdiction or Delaware. The nationwidesearch for attractive legal rules that Cary feared and Winter cheered 16

does not appear. 17 The nationwide race or market may be a heuristicfor potential competition, but it does not describe firm choices duringthe period studied. Thus, the dominant metaphor of a national racebetween fifty states or a single market with fifty producers is incorrectand potentially misleading.

This bimodal character of firm choice and the "home-state" biasis unexplained by the literature on state competition, which empha-sizes firm search for domicile based solely on legal rules.

Structure at the Initial Public Offering (May 2001), http://papers.ssrn.com/sol3/pa-pers.cfm?abstractid=236035 (describing venture financing and CEO-shareholder bargainas determinants of board structure). These studies have not, however, examined the roleof corporate law and legal protections for shareholders. This omission is striking as it iscorporate law rules that will govern, interpret, and enforce the shareholder rights the en-trepreneur designs. See also Barry D. Baysinger & Henry N. Butler, The Role of Corpo-rate Law in the Theory of the Firm, 28 J.L. & Econ. 179, 179 (1985) ("[S]tate corporationlaws ... differ and ... firms will select their state of incorporation adaptively."); GuhanSubramanian, The Influence of Antitakeover Statutes on Incorporation Choice: Evidenceon the "Race" Debate and Antitakeover Overreaching, 150 U. Pa. L. Rev. 1795 (2002).

14 But see Marcel Kahan & Ehud Kamar, The Myth of State Competition in CorporateLaw, 55 Stan. L. Rev. (forthcoming 2002) (manuscript on file with the New York UniversityLaw Review) (analyzing state efforts to attract incorporations of public companies).

15 See infra tbl.4, weighted by assets. Home state is defined as the location of the firm'sexecutive offices.

16 See William L. Cary, Federalism and Corporate Law: Reflections Upon Delaware,83 Yale L.J. 663 (1974); Ralph K. Winter, Jr., State Law, Shareholder Protection, and theTheory of the Corporation, 6 J. Legal Stud. 251 (1977).

17 Note that potential competition remains. That is, though Delaware currently maynot face meaningful competition from other states, Delaware's behavior would be con-strained by the possibility that other jurisdictions might begin to compete if Delaware ne-glected firm demands.


[Vol. 77:1559


The evidence I find suggests that the question with which thestate-competition literature has been primarily concered-"arestates racing to the top or bottom?"-may not be the right question atthis point. Instead of offering theoretical accounts to believe that therace is up or down, we should also ask: (1) What explains the gravita-tional pull of the home state? (2) How do individual firms select astate of incorporation? (3) What factors lead to a state's success inattracting firms? (Do takeover laws attract firms? Does corporatelaw have network benefits? Is the Model Act 8 valuable?)

In this Article, I offer initial and tentative answers to these ques-tions (see Table 1 below for a summary). But I take a central contri-bution of this Article to be that it begins to examine the "industrialorganization" of the IPO market and firm demand for corporate lawrules.19

TABLE 1

Standard assumptions about the The actual marketrace to the top/bottom For corporate lawThere is a nationwide race between states Firms choose either Delaware or their homeand entrepreneurs search out and choose state. Delaware engages in serial competi-between fifty alternatives. tion in local markets but does not face

actual competition in the market fornational firms.

Delaware is winning the race with a roughly Delaware's has a nearly 70% share of IPO50% market share. firms and 95% share of firms incorporating

outside their home state. Delaware's shareis growing over time.

A state's market share depends solely on its The location of a firm's executive officeslaws and firms choose incorporation state greatly affects the firm's incorporation,based on law. probably because lawyers matter.

Managers or shareholders choose corporate Lawyers influence incorporation choices andlaw rules. may seek their own welfare.

Part I describes the debate over the market for corporate law.Part II describes the Article's methodology and the sample of IPOfirms. Part III presents the unique market structure for IPO firmcharters that I observe. Part IV examines possible reasons for ahome-state advantage. Part V tests several hypotheses about the de-mand for Delaware domicile, the type of laws that firms value, andwhy some firms incorporate in Delaware while others do not. I focuson the role of the lawyer in connection with the IPO firm's domicilechoice. Note, however, that I can make only tentative claims as to the

18 Model Bus. Corp. Act, at xxvii (1998/1999 Supp.).19 A separate paper will examine what these facts and serial competition imply for the

state-competition debate generally and will present additional statistical evidence. ThisArticle describes the market structure and presents preliminary regression results.



causes of a firm's choice of state of incorporation, as firm domicile isjointly determined with the firm's other financial and governancecharacteristics.

THE DEBATE ABOUT THE MARKET FOR CORPORATE LAW

Perhaps the most basic debate in corporate law concerns firms'freedom to select their state of incorporation (and with it, the corpo-rate law rules that govern the firm). This debate has generated in-tense disagreement and has influenced broader debates aboutfederalism and state regulation in areas such as securities regulation,bankruptcy, environmental regulation, and banking regulation.20 ThisPart argues that the state-competition debate has assumed a set ofstylized facts about firm choice of law that, in material ways, do notcorrespond to observed behavior.

A. The State-Competition Debate

Critics of state regulation claim that, free to choose, self-inter-ested managers seek lax or entrenching corporate laws in order toprofit at shareholders' expense, leading states to provide promanagerlaws in order to attract incorporations.21 Delaware, the most success-ful state on this score, is thus said to cater to managers by removingshareholder protections: "[A] pygmy among the 50 states prescribes,interprets, and indeed denigrates national corporate policy as an in-centive to encourage incorporation within its borders. '22 Critics thuscall for Congress to pass mandatory laws to regulate fiduciary duties,takeovers, and shareholder voting in order to protect investors.23

By contrast, supporters of state regulation argue that marketforces (such as the capital, labor, and product markets) limit such op-

20 The state-competition debate influences a wide variety of other areas. See, e.g.,Henry N. Butler & Jonathan R. Macey, The Myth of Competition in the Dual BankingSystem, 73 Cornell L. Rev. 677 (1988) (discussing theory of competition between state andfederal banking regulators); Robert Rasmussen, Resolving Transnational InsolvenciesThrough Private Ordering, 98 Mich. L. Rev. 2252, 2273 (2000) (arguing that regime ofbankruptcy selection clauses would encourage competition among nations for bankruptcybusiness and "would lead to a general increase in the quality of extant law"); Richard L.Revesz, Rehabilitating Interstate Competition: Rethinking the "Race-to-the-Bottom" Ra-tionale for Federal Environmental Regulation, 67 N.Y.U. L. Rev. 1210 (1992) (discussingstate-competition debate with respect to environmental regulations); Roberta Romano,Empowering Investors: A Market Approach to Securities Regulation, 107 Yale L.J. 2359,2365-2401 (1998) (advocating market-oriented approach that would expand role of statesin securities regulation).

21 See Cary, supra note 16 (advocating increased federal role in corporate law to ad-dress deteriorating corporate standards resulting from "race to the bottom").

22 Id. at 701.23 See, e.g., Bebchuk, supra note 9.


[Vol. 77:1559


portunism and that managers select, and states therefore compete toprovide, rules that increase firm value.24 Therefore, "entrepreneurschoosing a state of incorporation will search for legal rules that maxi-mize investor's welfare," and state competition leads to a national"race to the top."'' 5 Because firms voluntarily seek out valuable inves-tor protections and because they benefit from having a variety of legalregimes from which to choose, mandatory corporate law rules arelikely to be harmful.26

B. Common Assumptions About Firm Choice

The two camps thus disagree about whether state competitionbenefits shareholders or managers-but these positions are otherwisevirtually indistinguishable. Each camp adopts a common a priori viewof the nature of state competition and the market for corporate law.First, each side agrees that a firm's choice of incorporation state isentirely decoupled from its operating decisions. Firms may, withoutcost or special permission, operate in State A and incorporate in StateB. The choices are considered entirely independent. 27

Second, each side agrees that legal rules affect the creation ordistribution of the firm's value and that firms therefore search the ar-ray of fifty potential legal regimes and select the regime with the mostfavorable legal rules. Thus, Easterbrook and Fischel claim that man-agers and investors search out the best corporate law from among thefifty competing states.28 They caution, however, that "[t]here are onlyfifty states, perhaps too few to offer the complete menu of termsneeded for the thousands of different corporate ventures .... [W]ithonly fifty jurisdictions, lots of (potentially desirable) combinations of[laws] will be missing. '29

24 See Winter, supra note 16 (rejecting "race for the bottom" theory and arguing thatcompetition aligns management's interest with shareholders).

25 Easterbrook & Fischel, supra note 4, at 213.26 Roberta Romano, The Genius of American Corporate Law 91-117 (1993) (examin-

ing impact of national laws regulating mandatory disclosure, insider trading, and federalcriminalization of state fiduciary duty).

27 Under a long-standing "internal affairs" rule, the corporate law rules of the incorpo-rating state govern the firm's internal disputes. See supra note 6. Thus, the only conse-quence of choosing a state of incorporation is the determination of the corporate law rulesthat will govern the firm.

28 Easterbrook & Fischel, supra note 4, at 213-15.29 Id. at 216 (emphasis added). Macey and Miller note that "a decision about where to

incorporate [is difficult because it] involves a comparison of legal rules across fifty jurisdic-tions." Jonathan R. Macey & Geoffrey P. Miller, Toward an Interest-Group Theory ofDelaware Corporate Law, 65 Tex. L. Rev. 469, 487 (1987). Posner and Scott suggest thatfirms search out states that specialize in providing law for their particular type of firm.Richard A. Posner & Kenneth E. Scott, Economics of Corporation Law and SecuritiesRegulation 111 (1980) (speculating that Delaware has tailored its law towards needs of



Third, both camps also agree that a state's market share is deter-mined by the substance of its legal rules. In the words of one state-competition critic, "According to both the race for the bottom andrace for the top theories, states' performance in the competition forincorporations is determined by the substantive content of their legalrules.' 30 That is, both camps assume that State A will attract 10% ofthe nation's firms if its corporate law rules are offered at a price at-tractive to 10% of the nation's firms.

Fourth, both sides agree that states compete to provide legal rulesthat firms want because incorporation leads to incorporation fees andwork for local service providers, such as lawyers. 31 Delaware, hometo roughly 50% of the Fortune 500 firms, has proven so successful thatfees from incorporations make up more than 20% of the state's reve-nues.32 Other states are also said to compete for this incorporationbusiness.33 Nevada, for instance, is frequently described as trying tobecome the "Delaware of the West," and other states are thought tocompete by adopting antitakeover measures that favor managers.34

The fact that firms have a choice have led many to analyze corpo-rate law as a product or to conceive of this as a nationwide race toprovide corporate law.35 These analogies have proven exceedinglypowerful and most scholars today either explicitly or implicitly accepta version of state competition that is described more generally as Tie-bout 36 competition: Firms select valuable legal rules from a nation-wide menu of alternatives.37

large, public corporations, while other states have targeted small, closely heldcorporations).

30 Bebchuk, supra note 9, at 1446 (emphasis added).31 See Kahan & Kamar, supra note 14 (manuscript at 11-35) (analyzing state incentives

to compete for incorporations).32 Marcel Kahan & Ehud Kamar, Price Discrimination in the Market for Corporate

Law, 86 Cornell L. Rev. 1205, 1211, 1225, 1251 (2001).33 Some scholars claim other states such as Maryland attract a substantial number of

firms, suggesting success in competition with Delaware. See Subramanian, supra note 13 at1816 (showing that Maryland holds second largest share of U.S. public company incorpora-tions); see also Lucian Bebchuk & Alma Cohen, Firms' Decisions Where to Incorporate 3(January 2002) (unpublished manuscript, on file with the New York University Law Re-view) (stating that Maryland has "significant toehold[ ] in the market for out-of-stateincorporations").

34 Kahan & Kamar, supra note 14 (manuscript at 64). For summaries of state antitake-over statutes, see John H. Matheson & Brent A. Olson, Shareholder Rights and LegislativeWrongs: Toward Balanced Takeover Legislation, 59 Geo. Wash. L. Rev. 1425, 1519-69(1991).

35 See, e.g., Cary, supra note 16; Winter, supra note 16.36 See Charles M. Tiebout, A Pure Theory of Local Public Expenditures, 64 J. Pol.

Econ. 416 (1956).37 See, e.g., Daniels, supra note 9, at 138-40 (1991).


[Vol. 77:1559


Given these assumptions, one would expect to find dispersion infirm incorporation choices. If State A offered a superior (inferior) setof rules, firms would migrate to State A, regardless of their headquar-ters. State B might respond by offering an even better (worse) set ofrules, so firms would then move to State B. However, because bothmanagers and shareholders have a veto power over reincorporations,their interests may diverge and presumably some firms would remainin State A. The result would be some national distribution of firmdomicile where a state's market share reflected the past and currentvalue of its legal rules.38 A similar dispersion of firm incorporationswould result if states specialized in providing laws for particular sortsof firms.

Because so little is known about the structure of this market andthe degree to which reality matches the stylized facts, this Article in-vestigates the market for IPO charters to examine what firms actuallydo.

C. The Importance of Understanding the Market forCorporate Law

In addition to shedding light on the state-competition debate, un-derstanding how firms choose corporate law is valuable for other rea-sons. First, it affects our understanding of earlier findings thatDelaware firms were more valuable. If valuable firms and futuretargets simply search out Delaware, this suggests that Delaware doesnot add value.39 Second, choice of domicile is important as a practicalmatter. In the past twenty years, more than six thousand firms havegone public, at a rate of more than one a day, raising $488 billion (in

38 This predicted pattern of flipping and national dispersion of domicile posed a prob-

lem for the original race-to-the-top/bottom debate. Romano noted that the standard ac-count implies flipping and dispersion, supra note 13, at 228. Posner & Scott similarlysuggest that the market is characterized by product differentiation: States offer differentpackages of investor protections, allowing firms to sort themselves into the appropriatedomicile, supra note 29. Romano offers an alternative theory (and evidence) for Dela-ware's sustained lead and leaves the incorporation in other states unconstrained, whichunder the standard assumptions of the competition hypothesis may lead to dispersal. Onepossible explanation, explored here, is that there is a home state advantage or bias that isimportant relative to existing variation between state corporate laws. See text, infra at1576 and 1602.

39 If valuable firms and likely takeover targets simply prefer to incorporate in Dela-ware, this would raise interesting questions about why Delaware is attractive to these firmsand it would help explain Delaware's advantage. But states would have no special reasonto imitate Delaware law generally. On the other hand, if Delaware helps to create-ratherthan merely attract-valuable firms, this suggests that Delaware law is relatively valuableand that firms generally incorporate in a value-improving jurisdiction.



2001 dollars) in gross proceeds. 40 If IPO firms, on average, sell 30%of their stock at the IPO, the market value of these firms at the end ofthe first day's trading would exceed $1,900,000,000. If corporate lawaffects the value, productivity, or internal organization of these firms,the question of how firms select domicile is important.

Third, understanding how firms select corporate law rules mayaffect policy debates concerning the regulation of shareholder rightsaround the world. For instance, the European Union (EU) may adopta similar U.S.-style rule that allows firms to choose the corporate lawregime of any Member State.41 The debate over whether the EUshould open shareholder rights up to freedom of choice draws heavilyon interpretations of the U.S. experience with regulatory competition.

IIMETHODOLOGY AND SAMPLE

A. Why Choices of IPO Firms Matter

The few previous studies of domicile choice have examined thereincorporation choices of public firms (the canonical reference isRoberta Romano's classic study). 42 These studies are quite useful andhighlight the reasons firms change domicile. However, by definition,studies of reincorporating firms do not focus on firms that stay put;many firms never reincorporate and the vast majority do notreincorporate after they are public. 43 Moreover, because the subset ofreincorporating firms is not a random sample of public firms, it is hardto generalize the results. Domicile decisions of reincorporating maybe fundamentally different than decisions of other firms. By examin-ing the incorporation choices of all IPO firms (whether or not they

40 This figure is obtained from Securities Data Corporation (SDC) data and includesthe average increase in stock price of 18.8% on the first day of trading, which declines onaverage in following years. Jay Ritter & Ivo Welch, A Review of IPO Activity, Pricing,and Allocations, 57 J. Fin. 1795, 1797 tbl.1 (2002).

41 A recent decision by the European Court of Justice has raised the possibility that theEU may abandon the "real seat rule"-the rule that firms are subject to the corporate lawrules of the country in which they operate-in favor of something like the U.S. "internalaffairs rule"-which is that firms are free to incorporate anywhere and their internal affairsare subject to the corporate law rules of the incorporating jurisdiction. See Case C-212/97,Centros Ltd. v. Erhvervs-og Selskabsstyrelsen, 1999 E.C.R. 1-1459, [1999] 2 C.M.L.R. 551(finding contrary to European Communities Treaty Danish government's refusal to regis-ter branch of company which conducted no business in United Kingdom but was incorpo-rated there in order to avoid Danish laws on minimum capital requirements). The ultimateeffect of the Centros decision is uncertain. See e.g., Eddie Wymeersch, Centros: ALandmark Decision in European Company Law (Fin. Law Inst., Working Paper Series No.1999015, 1999), available at http://papers.ssrn.comlsol3/papers.cfm?abstractid=190431.

42 See Romano, supra note 13; Subramanian, supra note 13, at 1795.43 Daines, supra note 7, at 527.


[Vol. 77:1559

December 2002] INCORPORATION CHOICES OF 1PO FIRMS

reincorporate), I can examine all firms' domicile choices at a time thatfirms commonly evaluate and change their domicile.

IPO incorporation choices are also interesting because, in con-trast to reincorporating public firms, they are relatively free from theagency costs. Managers of firms that are already public typically ownlittle stock and therefore do not bear the full cost or benefit of thefirm's legal rules. As a result, they may prefer either to remain in ormove to a state with inefficient legal rules: A small increase in theirperks or private benefits may outweigh their share of the resultingdrop in the firm's stock price. Thus, agency costs may lead publicfirms reincorporating midstream to demand value-reducing legalrules.

However, while public firms may make suboptimal domicile deci-sions, it is less likely that firms going public will search out bad rules.IPO firms make domicile decisions before ownership is dispersed andbefore collective action problems and agency costs become severe.Before an IPO, managers own more than 50% of the firm's stock44

and thus bear much of the cost of any suboptimal decision. Moreover,other pre-IPO investors (such as repeat players like venture capital-ists) also have large pre-IPO holdings and have an incentive to moni-tor managers.

IPO domicile choices are also important because they are likelyto shape the corporate law rules that states produce (and thus createthe race to the bottom/top). More than six thousand firms went publicbetween 1978 and 2000, and as part of this process were likely to eval-uate their domicile with some care. These choices are typically per-manent. By contrast, only roughly six hundred public firmsreincorporated in the same period. Thus, if states compete to producecorporate law rules that firms value, they are likely to take the prefer-ences of IPO firms into account.

Finally, incorporation decisions at the IPO stage shed light onhow (and to what degree) entrepreneurs use corporate law rules toregulate investors' rights and limit agency costs at the firm'sinception.45

Therefore, although there is evidence about how and why publicfirms reincorporate, it is useful to examine the charter market for IPOfirms.

44 See Daines & Klausner, supra note 1, at 93 tbl.1.45 See generally Baker & Gompers, supra note 13; Daines & Klausner, supra note 1;

Kaplan & Stromberg, supra note 4; Karpoff & Field, supra note 13.



B. Sample Description

I examine a sample of 6,671 IPOs between 1978 and 2000.46 Toselect the sample, I begin with all domestic firms that completed abest-efforts IPO in the time period, according to Securities Data Cor-poration (SDC) records. I then exclude firms that do not have finan-cial data available on Compustat. Mutual funds, financial firms andReal Estate Investment Trusts (REITs) are excluded from the samplebecause the law governing these firms differs substantially from thosegoverning industrial firms. Moreover, the central issues that occupycorporate law scholarship-takeover law and fiduciary duty in publicfirms-are less important in these firms given the mandatory state andfederal regulation to which they are subject.

Data on sample firms' current state of incorporation and the loca-tion of their headquarters come from Compustat.47 Data on mergerand acquisition activity, legal counsel, underwriters, auditors, and thefirms' headquarters come from SDC. Data on underwriter reputationcome from Carter and Manaster.48

The firms in the dataset represent a broad spectrum of firms fromall parts of the country and a wide variety of industries. The datasetincludes firms in 262 different three-digit Standard Industrial Classifi-cation (SIC) codes in each of the fifty states (and the District of Co-lumbia). Roughly 40% of the sample firms are located in California,New York or Texas. Though many firms are taken public by familiesor entrepreneurs, 11% are spin-offs (that is, former subsidiaries ofother companies) and roughly 42% had prior investment from venturecapitalists.

46 Regression estimates examine only firms that went public from 1978-1997, the lastyears for which Compustat data was available when I collected the data for the regressions.

47 Because Compustat only records a firm's current state of incorporation, it was neces-sary to identify firms that reincorporated since going public. Firms reincorporating since1987 were identified by a LEXIS search of proxies and 10-K filings of all public firms.Data on reincorporations between 1978 and 1987 came from Moody's Annual IndustrialReports, as collected in Romano, supra note 13, at 242, and Robert Comment & G. Wil-liam Schwert, Poison or Placebo?: Evidence on the Deterrence and Wealth Effects ofModern Antitakeover Measures, 39 J. Fin. Econ. 3, 11 (1995).

48 See Richard B. Carter & Steven Manaster, Initial Public Offerings and UnderwriterReputation, 45 J. Fin. 1045, 1056-66 (1990); Richard B. Carter, Frederick H. Dark & AjaiK. Singh, Underwriter Reputation, Initial Returns, and the Long-Run Performance of IPOStocks, 53 J. Fin. 285 (1998).


[Vol. 77:1559


IIITHEo MARKET FOR CORPORATE LAW

A. Overview

Consistent with the view of an active national market, corporatecharters appear, at first glance, to be widely distributed among states.Sample firms are incorporated in all fifty states. Although Delawarehad the largest market share (capturing over 50% of the IPOs, consis-tent with frequent reports that Delaware has a roughly 50% marketshare), firms still found other states' corporate laws attractive. Cali-fornia was home to roughly 8% of all IPOs, while New York, Massa-chusetts, Florida, Minnesota, and Texas collectively capturedapproximately 14% of the IPO market. This is summarized in Table 2below and additional detail can be found in Appendix A.

TABLE 2MARKET SHARE OF IPO CHARTERS

State Market Share

Delaware 56.0%California 8.4New York 3.2Minnesota 2.9Florida 2.7Massachusetts 2.7Texas 2.6New Jersey 1.8Pennsylvania 1.6Nevada 1.5Washington 1.4Ohio 1.3

By this measure, Delaware's share of the IPO market (and thusits advantage in producing legal rules) is large, but not overwhelm-ing.49 Buyers of corporate law appear to have alternative producersto choose from, and some states seem to have success competing forIPOs. This is consistent with the standard views of both the race-to-the-top and race-to-the-bottom advocates that there are numerouscompetitors, but Delaware leads with roughly 50% of the market.50

49 The Herfindahl-Hirschman Index for the IPO market in the 1990s was roughly 3000(Department of Justice (DOJ) guidelines for horizontal mergers suggest that an industry ishighly concentrated when its index is about 1800). See U.S. Dep't. of Just. & Fed. TradeComm'n, Horizontal Merger Guidelines § 1.51 (1992), available at http://wvw.usdoj.gov/atr/publiclguidelines/horiz_book/15.html.

50 Because SDC data on IPOs only goes back to 1978, I hand collected data on NewYork Stock Exchange (NYSE) IPOs between 1960-1978, though data for year 1966 is miss-



Delaware's share of New York Stock Exchange (NYSE) IPOs has in-creased (from approximately 30% during 1960 to 1978, to 54% be-tween 1978 and 1998). See Table 3 below.

TABLE 3DELAWARE'S SHARE OF ALL IPOs LISTED ON NYSE

1960- 1965- 1970- 1975- 1980- 1985- 1990- 1995-Year 1964 1969 1974 1979 1984 1989 1994 1998

Delaware's share of IPOs 30% 30% 31% 36% 56% 75% 73% 77%Total number of IPOs 265 279 331 140 95 160 307 217

B. The Bimodal Market and Serial Competition

The image of firms making unfettered choices from a nationwidemenu of legal regimes disappears upon further examination. Firmsessentially incorporate in one of only two places: their home state orDelaware. Roughly 95% of all firms that seek legal rules outside theirstate headquarters end up in Delaware-a market share that wouldmake even Microsoft envious. Only a trivial fraction of firms seekout-of-state legal rules anywhere else: Delaware law governs morethan 97% of assets of public firms incorporated outside of their homestate, while together, Delaware's nearest four rivals split 0.6% of themarket between them. See Table 4 and Figure 2 below and AppendixA for additional detail. 51

ing. Though the increase appears dramatic, one must be cautious in interpreting any over-all trend because data for the first half of the period come from a different source (NYSE)that may have had different data collection and sample criteria. I am grateful to TaraBenzinger for assistance in collecting these data.

51 Subramanian reports a greater percentage of firms reincorporating in third states.Subramanian, supra note 13, at 1825-26. There are likely several reasons for this. First, hissample includes REITs and financial firms, which I also find are likely to incorporate inother states (typically Maryland, Massachusetts, or Minnesota). I however exclude thesefirms and focus on the industrial firms on which the state-competition debate has centered.The law governing REITs and investment companies is a specialized area of law and firmchoice is not based on state regulations of the beneficiary-manager relationship, which isgoverned by federal law. See Kahan & Kamar, supra note 14 (manuscript at 61). Moreo-ver, many firms that Compustat records as out-of-state firms incorporated in Massachu-setts (or Maryland), for instance, are not companies at all, but business trusts organized tocreate mutual funds. Alternatively, it may also be that reincorporating firms are signifi-cantly different from IPO firms or that some of the firms reincorporate to a third state aspart of a merger with another operating firm.


1572 [Vol. 77:1559


TABLE 4LEADERS IN THE RACE TO ATTRACT OUT-OF-STATE FIRMS

Market Share Market ShareState (by number of firms) (by total IPO firm assets)

Delaware 94.8% 97.0%Nevada 1.5 0.3California 0.7 0.2New York 0.7 0.04Minnesota 0.6 0.1

/ '--- "-,'" -DE -out of state/',,,J --- DE

1980 20001990

FIuRE 1

IPO Market Share: 1980-89 IPO Market Share: 1990-2000

E3DE U DEEHome U Home0 the 0ther

FIGuRE 2

Thus, no state besides Delaware has had any meaningful successin attracting out-of-state firms going public. Other states attracted onaverage only three out-of-state firms and the median state attractedonly one out-of-state firm over the sample period.5 2 Even Nevada,

52 In unreported tests, I find that the vast majority of Maryland incorporations are mu-tual funds, REITs, and investment companies. Maryland's ability to attract such firms isnot related to its corporate law rules and regulation of the shareholder-manager relation-ship but is instead a function of its rules governing authorized shares. Kahan & Kamar,supra note 14 (manuscript at 70). Thus, Maryland does not provide an example of a stateable to compete on the central aspect of corporate law-regulation of the fiduciary dutiesof managers and shareholder protection. Cf. Bebchuk & Cohen, supra note 33.



ostensibly the most aggressive competitor, has had only trivial suc-cess-attracting only thirty-four out-of-state firms of the roughly 3400that went public between 1990 and 1997 (or 0.3% of assets).5 3 Thus,while Delaware's share of IPOs is large, its share of firms that decideto go out of state for legal rules is overwhelming.5 4

Although the standard assumption is that legal domicile has nonecessary connection to a firm's headquarters, the gravitational pull ofa firm's home state is so strong that only about 5% of all firms thatincorporate outside of Delaware make it anywhere else. The bimodalstructure of firm choice is illustrated in Figure 2 above, which exam-ines the market share of all IPOs between 1980-2000.

The financial consequences to states of attracting out-of-statefirms are trivial: from each of the thirty-four out-of-state firms incor-porated within its borders, Nevada collected annual franchise (or re-port) taxes of roughly $85 a year which totaled no more than $3000annually. These firms also paid relatively minor one-time fees basedon the aggregate authorized par value of capital stock. While a com-plete calculation of Nevada's gains would include legal fees paid bythese firms and certain filing fees for recording corporate events, theseare unlikely to be large.5 5 Consequently, it is hard to see how Ne-vada's oft-cited effort to become the "Delaware of the West" has pro-duced any financial gain, at least from public companies.5 6 Otherstates most "successful" in attracting out-of-state firms-California,New York, and Minnesota-also had no meaningful financial gainfrom their success: They collect trivial initial fees and imposefranchise taxes, based not on domicile but on the business done in-state. Thus the incorporations have no meaningful impact on thestate's marginal revenue.5 7

53 Nevada attracted only twenty-four nonfinancial firms over the whole decade of the1990s.

54 Because California firms are such a big part of the IPO sample, I also examine Cali-fornia's market share. California fares poorly (retaining 20 to 30% of local firms), andsome of Delaware's increase may simply be a mechanical relationship-as more firms fromCalifornia go public, Delaware's share will improve.

55 See Kahan & Kamar, supra note 14 (manuscript at 21, 29-30).56 Indeed, it appears that Nevada, to the extent it competes for incorporations, does not

target the market for public firms. Nevada law firms and business-service firms emphasizequalities of Nevada law that are unlikely to appeal to public shareholders or underwriters,such as the fact that Nevada does not share tax information with the IRS. See, e.g., Incor-porate in Nevada: Why Nevada Corporations Are Popular for Business, http:llwww.mycorporation.comvhynevada.htm (last visited Nov. 8, 2002); Fortress Protection:Nevada Corporations, http:/lwww.angelfire.com/ca7/fortress/nevada.htm (last visited Nov.8, 2002). Moreover, during this sample period Nevada did not even have a specializedcorporate court, something one would expect from a jurisdiction intent on attracting publicincorporations.

57 See Kahan & Kamar, supra note 14 (manuscript at 14-15).


[Vol. 77:1559


C. Discussion

Free to incorporate anywhere, to search out exotic legal regimes,the overwhelming majority of firms select either their home state orDelaware. This bimodal pattern is perhaps the most basic characteris-tic of the market for corporate charters-it describes roughly 95% ofdomicile decisions. Any useful account of state competition or corpo-rate law must explain this pattern.

These facts stand in marked contrast to the image of nationwidecompetition for firms evoked by the metaphor of a national race be-tween fifty states. Firms do not act as if they are selecting from anationwide menu of potential legal regimes. There is not nationwideTiebout competition in any meaningful sense. 58 Rather, there are ap-parently forty-nine separate markets or linked duopolies in whichDelaware engages in serial competition with local producers-com-peting with New York for New York firms, with California for Califor-nia firms, and so on. Thus, the central metaphors for state productionof corporate law - a nationwide race between 50 states or a singlemarket with 50 producers - is inaccurate or inadequately specified.

The balance of this Article briefly addresses questions raised bythe unusual, bimodal character of the market for corporate law rules:(1) What explains the strong gravitational pull of the home state andthis odd equilibrium? The existing literature does not explain such abias, although it is strong enough to account for roughly 95% of allincorporation choices. (2) How do individual firms select a legal re-gime? and (3) What explains variation in states' ability to competewith Delaware?

5S A defender of Tiebout competition would argue that the sorting occurs a step earlier:Firms locate in the state with optimal legal rules. This explanation seems implausible.First, very few firms locate business facilities in Delaware, a state with obviously desirablelaws. This is because there are cheaper ways to get the laws of State A than moving opera-tions to State A-incorporating in State A. Second, there is evidence that even environ-mental regulation (which is likely to affect firm costs) does not affect decisions aboutwhere to locate plants. Instead, firms choose based on factors like labor costs and educa-tion. See Revesz, supra note 20, at 1235 (positing that firm-location decisions are based onmultiple complex factors, only one of which is local environmental regulations); see alsoAdam B. Jaffe et al., Environmental Regulation and the Competitiveness of U.S. Manufac-turing: What Does the Evidence Tell Us?, 33 J. Econ. Lit. 132, 158 (1995) (concluding thatenvironmental regulations do not have significant effect on competitiveness of manufactur-ers and that "other factors would indeed overwhelm the environmental effect").



IVTHE HOME-STATE ADVANTAGE

A. The Existence of a Home-State Bias or Advantage

The pattern of choice is inconsistent with the idea that firms careonly about corporate law rules. If firms care only about corporatelaw, and if there is vigorous competition and variation between states,one would expect more migration between non-Delaware states. Forinstance, California produces corporate law rules good enough to at-tract roughly 500 California-based IPOs-but not a single firm fromNew York or Pennsylvania. It is implausible that California corporatelaw rules are good enough to attract this many California firms butnot good enough for outsiders. Consider also Oregon: Its record atretaining its own firms is excellent-tops in the nation for the pasttwenty years, retaining almost 70% of its IPO firms. 59 But Oregon isattractive only to Oregon firms. Only three out-of-state firms incorpo-rated in Oregon during the two decades I examined.60

Why don't firms in Idaho (which retained none of the Idaho firmsthat went public-a 0% retention rate), New York (12% retentionrate), or Illinois (12% retention rate) migrate to Oregon (69% reten-tion rate), Minnesota (65% retention rate), or Louisiana (53% reten-tion rate)?61 In short, why don't firms migrate away from low-retention/bad-law states and into high-retention/good-law states?How can corporate law affect the domicile's value to local firms andthus a state's ability to retain local firms, and why doesn't this varia-tion attract foreign firms? Note at the outset that it is no answer toreply that states do not compete - the question is why firms don'tmigrate, whatever the source of apparent variation in states.

One answer may be that local incorporation brings benefits otherstates cannot match. Imagine that states offer incorporating firms twothings: corporate law rules (worth $2 if the state's law is good for thefirm, and $2 if the law is bad) and the value of a home-state advantage(worth $10). Assume that Delaware law is uniquely valuable (worth$9). In this simple example, if home state law is good for a firm, thefirm will incorporate at home (home incorporation will be worth $12,

59 For more complete data on states' differing ability to retain local firms, see infraapp.A tbl.A3.

60 Interestingly, two of these three firms were from Oregon's next door neighbor,Washington This is an example of a general pattern, where firms incorporating outside oftheir own state or Delaware are more likely to incorporate in an adjacent state. See infratbl.8 and accompanying text. This again suggests the importance of physical proximity.

61 There appears to be meaningful variation in states' ability to attract their local firms.In the 1990s, ten states retained 50% or more of the IPO firms located in-state, whilefourteen retained fewer than 15% and seven retained none. See infra app.A tbl.A2.


1576 [Vol. 77:1559


Delaware $9, and all other states either $2 or -$2). If the home statelaw is bad, the firm will incorporate in Delaware (Delaware will beworth $9, home state worth $8, and all other states worth either $2 or -$2). Thus, in this simple example, if the home state advantage is verylarge (relative to the differences between states) variation in local lawcould affect the number of firms incorporated at home, but would notdrive firms to jurisdictions other than Delaware. Other jurisdictionswouldn't be able offer corporate-law benefits to match Delaware andcould not (by definition) offer a home-state advantage.

But what home benefits or biases might there be? The existingliterature does not provide an explanation. This Section considersseveral possible explanations, including transaction costs, imperfectinformation, and agency costs.

Before considering these explanations, however, I note that theidea of a home bias may strike some as implausible: IPO entrepre-neurs have incentives and freedom to select valuable rules and in-termediaries can search legal rules of alternative states. Meaningfulvariation would thus be expected to produce migration. It is thereforeuseful to note that are similar biases have been found in other mar-kets, whether information is even more transparent and the stakes areeven higher. Though new in corporate law, such a home bias is "per-haps the least controversial stylized fact"62 in international capitalmarkets, where stakes are at least as high and actors are at least aswell-informed as the legal market. In general, the home bias in thefinance literature refers to the fact that equity investors overinvest intheir local markets. If investors cared only about the risk and returnof their portfolios, as theory suggests, they would hold the world port-folio and invest proportionally in each country. Instead, investorsconsistently invest too much in local markets. For example, U.S. in-vestors have 91% of their equity investments in U.S. firms, eventhough U.S. firms represent only 49% of the world market portfolio.63

This home bias persists even absent significant transaction costs, regu-latory barriers or accounting differences and even though investorshave strong incentives to seek optimal returns; thus, a large literaturehas not succeeded in providing an explanation for the phenomenon.64

If such a bias persists in markets as thick and liquid as public equity

62 Lee Pinkowitz et al., Corporate Governance and the Home Bias, Social Science Re-

search Network, at http:llpapers.ssrn.comlsol3/papers.cfm?abstract id=292835#Paper%20Download (Nov. 2001).

63 Id. at 1.64 A similar home bias occurs in a nation's consumption patterns. See generally Karen

Lewis, Trying To Explain Home Bias in Equities and Consumption?, 37(2) J. Econ. Lit.,571 (1999).



markets, its seems reasonable to inquire if such a bias matters to in-corporation decisions.

B. .What Explains the Home Bias?

A useful account of state competition or corporate law must ex-plain the most basic features of firm choice of corporate law: thegravitational pull 65 of the home state and the bimodal character offirm choice. This Section takes initial steps at identifying two possibleexplanations for the observed home bias or home court advantage:The bias may be a function of regulatory benefits granted local firmsand may be partly a function of the market for legal services. Eachmay identify benefits to local firms that are unrelated to the state'stechnical corporate law rules.

1. Political Influence

The only legal consequence of incorporation choice is that it spec-ifies the corporate law that will apply. Zoning, environmental, labor,or tax benefits all apply to firms that do business in-state, regardless ofwhether the firm is incorporated. 66 The most plausible benefit fromlocal incorporation is that it allows managers to influence the firm'scorporate law rules by lobbying the legislature for particular corporatelaw reforms. When corporate rules constrain managerial action or af-fect the timing, cost, or feasibility of important transactions or govern-ance changes, the firm's managers may lobby the legislature to changethe rules.

The firm's ability to obtain this relief is likely to increase with thecontacts between the firm and the state, which might lead to an ob-served home bias. A firm with local productive assets, a large em-ployee base, and a traditional association between the firm and thecommunity is better positioned to lobby for corporate law reform. If a

65 Gravity may be an appropriate model for the flow of incorporation between statesand has been used to model migration patterns and trade flow. Newton's Law of UniversalGravitation states that two bodies in the universe attract each other in proportion to theproduct of their masses and inversely to the square of the distance between them. Migra-tion flows are sometimes analyzed with a gravity model-other things equal, there will bemore migration to a city that is (1) close and (2) large. George Kingsley Zipf, The PjPzIDHypothesis: On the Intercity Movement of Persons, 11 Am. Soc. Rev. 677 (1946). There issome evidence that firms act in this way: They are more likely to incorporate nearby andin a state that is larger (home to more corporations).

66 Moreover, since franchise fees for a particular firm are typically trivial-and for moststates unrelated to the firm's domicile-regulators are not likely to ask for local incorpora-tion as a condition of granting some requests.


[Vol. 77:1559


firm incorporates out of state, these assets and local connections areless valuable.67

Note, however, that the ability to influence corporate law rules isa double-edged sword: Managers deploy the firm's resources, includ-ing its political resources, but they may use their influence to benefitthemselves rather than the firm generally. In state after state, manag-ers have successfully lobbied for statutes that restrict takeovers andprotect managers (and possibly employees) at the expense of share-holders.68 For instance, faced with a takeover bid, Greyhound soughtprotective legislation from the Arizona state legislature. Given thefirm's number of employees and the importance to the state, in thewords of one legislator, "Greyhound said 'Jump' and we said, 'howhigh.' "69

Incorporation choice is thus one way to regulate the balance ofpower between managers and shareholders. Because managers havethe ability and incentive to seek protective measures from the legisla-ture, incorporating at home may be in some respects akin to a take-over defense - because the ability to lobby for reform is most likely tobe important in the takeover context.70

By contrast, as I have argued elsewhere, when firms incorporateout of state, where managers and employees lack local political clout,

67 Strictly speaking, managers might also influence legislators in states where they haveno operations. They may simply use capital to obtain local clout (via campaign contribu-tions and lobbyists). However, it will often be difficult to identify the state's interest inmaking the reform, given that for most states franchise fees (and the threat of the firm'sexit) may be trivial. Moreover, local firms can also hire lobbyists and make contributions,plus they have the additional asset of a local employee base (future voters) and local con-tacts that may help secure legislators' attention. This alone may provide a reason to incor-porate locally.

6S See Roberta Romano, The Political Economy of Takeover Statutes, 73 Va. L. Rev.111, 122-37 (1987) (discussing adoption of antitakeover statutes in Connecticut, Maine,Pennsylvania, Illinois, and Missouri driven by lobbying efforts of prominent local corpora-tions under pressure of hostile bids). This does not require the assumption that such lawsactually harmed shareholders but that legislatures respond typically to instate (potentiallyvoting) organized employees more than out-of-state (nonvoting) dispersed shareholders.Currently forty-two states have adopted one or more antitakeover statutes. Investor Re-sponsibility Research Ctr., State Takeover Laws 3 (1998); Lucian Bebchuk & Alma Cohen,Firms' Decisions Where to Incorporate, at A-9 tbl.9 (John M. Olin Ctr. for L. Econ. &Bus., Harvard Law Sch., Discussion Paper No. 351, July 2002), http://papers.ssrn.com/abstract=296492.

69 See Mark J. Roe, Takeover Politics, in The Deal Decade: What Takeovers andLeveraged Buyouts Mean for Corporate Governance 321, 339 (Margaret M. Blair ed.,1993) (quoting Arizona representative Jim Shelly).

70 For example, to fight off an attempted takeover, Norton Industries was able to se-cure a law classifying the board of all publicly traded firms incorporated in Massachusetts.Norton was a large employer and rallied its employees to spirited protest meetings. SeeRobert Daines, Do Classified Boards Affect Firm Value?: Takeover Defenses After thePoison Pill 16-20 (2001) (unpublished manuscript, on file with author).


1579


shareholders tie the hands of future managers and employees andmake it harder for managers and employees to extract protective an-titakeover legislation.71

2. Legal Advice

In considering incorporation decisions, we need to ask whomakes the decision, what information they have as they make it, andwhat incentives they have to invest in 'additional information. Thesequestions lead us to consider the role of lawyers and the possibilitythat the legal market influences incorporation choices and may helpexplain a home bias. This explanation is analogous to arguments thatthat imperfect information leads to a home bias in international capi-tal markets.72 If information about alternative legal regimes is costlyto acquire, then a home bias may result: Lawyers may not learn thecorporate law of other states, and entrepreneurs may not be able tomonitor or verify their lawyer's recommendation about domicile, per-haps giving lawyers some slack to seek their own, rather than the cli-ents' interest.

First, lawyers are likely to be influential in the firm's incorpora-tion decision, as they possess specialized legal knowledge and mayhave more at stake in the decision than other advisors. 73 For instance,a survey by Roberta Romano found that lawyers initiatedreincorporation decisions roughly 70% of the time and were almostfour times as likely to suggest reincorporating to Delaware than weremanagers. 74

Second, firms often hire local or in-state lawyers (more than 60%of sample firms did). This may be partly because the firm has alreadyretained local counsel to represent them before local courts, andwould incur switching costs if it changed legal counsel. Moreover, lo-cal lawyers have certain advantages. There is evidence that physicalproximity reduces informational asymmetry in business settings, par-ticularly "soft" information that is difficult to capture in written docu-ments.75 Firms also may have better information about the reputation

71 See Daines, supra note 7 (noting that Delaware has adopted relatively mild barriersto hostile takeovers and that few of its incorporated firms are targets with local politicalclout).

72 See, e.g., Joshua Coval & Tobias Moskowitz, Home Bias at Home: Local EquityPreference in Domestic Portfolios, 1999 J. of Fin. 2045; Lewis,supra note 64.

73 See Romano, supra note 13, at 273 (noting lawyers' role in suggesting that firmreincorporate).

74 Id. at 274 tbl.13.75 See, e.g., Coval & Moskowitz, supra note 72; Mitchell A. Petersen & Raghuram G.

Rajan, Does Distance Still Matter? The Information Revolution in Small Business Lend-ing, Social Science Research Networks, available at http://papers.ssm.comsol3/pa-


[Vol. 77:1559


of lawyers that operate in the same community. 76 These sorts of loca-tional and informational advantages are likely to be especially impor-tant for small (pre-IPO) firms, about which less is known and whereinformation asymmetries may be greater. Finally, if social or informalconnections are important in decisions about retaining counsel, thesefactors will lead firms to local lawyers.

If firms hire local lawyers, it is very likely that these lawyers, inturn, know local corporate law. State regulation requires lawyers topass local bar exams, which include questions about a state's corpo-rate law.77 Thus, many firms will hire local lawyers who know thecorporate law of the state in which they practice. Some lawyers alsoknow Delaware law, as it is something of a national standard, fre-quently taught in law schools. In particular, lawyers specializing inlarge, complicated transactions (where precedent and certainty is es-pecially valuable), or who have a multistate practice, are likely toknow Delaware law. Focusing on one national standard allows themto economize on the need to keep up to date with developments inmultiple jurisdictions. 78 Delaware is thus much like a common lan-guage and such lawyers are "bilingual," speaking Delaware law plusthe local dialect.

Thus, lawyers will know the law of the state in which they prac-tice, and some (particularly those with a multistate practice) will knowDelaware law. This would reinforce the bimodal pattern we observed.

Of course, lawyers may also learn the law of a third state, likeNevada. Network effects are likely to make this difficult, however.79

Learning another state's law is costly and a lawyer considering theinvestment vl consider the expected return, which is partly a func-

pers.cfm?abstract_id=223461 (Oct. 2000); Michael Storper & Anthony Venables, Buzz:The Economic Force of the City, available at http://www.druid.dklconferenceslsum-mer2002/Papers/STORPER.pdf (June 2002) (physical proximity can help overcome moralhazard problems and information flow and improve effort); Press the Flesh, Not the Key-board, Economist, Aug. 24, 2002 (physical proximity improves information flow and facili-tates trust and relationships necessary in deal making).

76 This is consistent with the fact that law firms typically open an office in a geographi-cal area in which they are likely to expand. Cumming and Macintosh report that reincor-porating firms in Canada prefer to have frequent face-to-face meetings with their counseland hire local lawyers to facilitate this face-to-face interaction. See Douglas J. Cummingand Jeffrey G. Macintosh, The Rationales Underlying Reincorporation and Implication forCanadian Corporations Int. Rev. Law & Econ. (forthcoming, 2002), available at http://papers.ssm.com/sol3/papers.cfm?abstractid=218391).

77 See, e.g., Jud. Ct. Acts § 478 (Consol. 2002); see also Romano, supra note 13, at 275-76 (describing survey results that local lawyers knew local law and therefore saw no advan-tage to incorporation in another state); infra Part V.A.1.

78 Romano, supra note 13, at 274-75 (noting that outside counsel are likely to have cost-saving incentives for suggesting domicile relocation to Delaware).

79 See infra text accompanying notes 93-95.



tion of the number of firms that adopt that state's law. If a lawyerexpected many firms to move to Nevada, she would have incentives tolearn Nevada law. However, since a large-scale shift to Nevada is dif-ficult to coordinate, lawyers discount this possibility and have reducedincentives to learn Nevada law.80 Thus, though this Section is largelyspeculative, in such a legal market where legal knowledge is costly toacquire, where lawyers know the law of their home state and somealso know Delaware law, the observed bimodal pattern and pairedcompetition may emerge, driven in part by the legal market and statebar regulations.

C. Magnitudes

Even if these effects lead to a home-state bias or advantage,before we could comment on their welfare effects, we would need toknow the extent to which there is meaningful variation in the corpo-rate law of states other than Delaware. If there is minimal variationbetween states other than Delaware, little would turn on the fact thatfirms rarely incorporate in third states.

There is reason to think there may be meaningful variation instate corporate law. First, Romano notes variation in states' substan-tive codes and the speed with which they update their statutes to re-spond to innovation.81 Second, states vary considerably in their abilityto retain local firms and these qualities appear to be durable over timeand to be affected by legal rules that the state adopts8 2 For instance,several states appear to be consistently attractive to their local firms,while others appear to be consistently less attractive.

Also, to the degree that firms are influential in lobbying for par-ticular reform, laws may vary somewhat between states. Thus, take-over statutes vary between states, though these are less likely tomatter to IPO firms (who can freely opt into or out of the statute)than to public firms.

On the other hand, if the core of corporate law is fiduciary duty(as seems reasonable), state law outside of Delaware may be largelyfungible. Delaware is uniquely attractive to firms, and its law and pre-cedent are hard for other states to duplicate.8 3 Fiduciary duty rules

80 See Romano, supra note 13, at 278.81 See id. at 233-42. Moreover, other states may not be effective competitors, as others

have also noted. See generally Kahan & Kramer, supra note 14.82 See infra app.A tbl.A2 (providing retention rates and regression results). Thus, the

very fact that the magnitude of the home-state advantage varies from state to state suggeststhat the value of corporate law rules similarly is not uniform among states other thanDelaware.

83 Daines, supra note 7, at 540 (noting difficulty of replicating Delaware's courts andstore of precedent); Kahan & Kamar, supra note 14 (manuscript at 81-83) (suggesting that


1582 [Vol. 77:1559


TABLE 5STATE SUCCESS OR FAILURE IN ATTRACTING

OUT-OF-STATE IPO FIRMS

This table examines a state's success in attracting incorporations among a sample of IPO firmsbetween 1980-2000.

1980-1989 1990-1999Out-of-state firms

State incorporated in state %

DE 583 94.20

NV 4 0.65

NY 3 0.48

NJ 3 0.48

MD 3 0.48

PA 2 0.32

MN 2 0.32

IN 2 0.32

GA 2 0.32

CA 2 0.32

All others 13 2.08

Total 619 100%

Out-of-state firms

State incorporated in state %

DE 2,293 95.38

NV 24 1.00

MD 14 0.58

CA 9 0.37

NY 8 0.33

MN 8 0.33

PA 5 0.21

GA 5 0.21

VA 4 0.17

IL 4 0.17

All others 28 1.16

Total 2,404 100%

are difficult to reduce to simple statutory rules (given the nature ofopportunism and the difficulty of predicting and legislating to preventfuture conflicts) and are instead regulated by a body of precedent.Other states, lacking Delaware's established precedent, unique caseflow, and specialized courts, are likely less able to develop a distinc-tive and predictable body of case law. Thus, corporate law in otherstates may be largely uniform and the statutory differences relativelyminor compared to Delaware. Moreover, one cannot exclude the pos-sibility that variation in state retention levels is due to differences inthe type of firms going public, rather than the substantive law. Thus,though we observe a home bias, even if we could discern its causes,

copying Delaware statutory law "would not deliver the exact same product that Delawaredoes" because of Delaware's unique court system and corporate case law); Ehud Kamar, ARegulatory Competition Theory of Indeterminacy in Corporate Law, 98 Colum. L. Rev.1908, 1932-35 (1998); Michael Klausner, Corporations, Corporate Law, and Networks ofContracts, 81 Va. L. Rev. 757, 846 (1995) ("[W]ithout a base of corporate domiciliaries andhence a litigation flow comparable to Delaware's, it would be difficult for another state tooffer a corporate law judiciary consistently as good as, let alone better than, Delaware's.");Romano, supra note 13, at 277 (noting that Delaware's "substantial body of case law can-not be easily replicated by a competitor"). Moreover, Delaware's record and reliance onfees provides firms and investors with assurance that, if Delaware courts make decisionsthat stray far from the expectations of business planners, legislative correction will quicklyfollow. See, e.g., Del. Code Ann. tit. 8, § 102(b)(7) (2001), which the Delaware legislaturespeedily passed in response to Smith v. van Gorkom, 488 A.2d 858 (Del. 1985).



the evidence I present here is still consistent with the idea that Dela-ware is unique, but that differences between other states are relativelysmall.

The next Part examines the role of lawyers and the need for polit-ical influence in the incorporation decisions of individual firms.

VTHE DEMAND FOR DELAWARE LAW

This Part examines the possibility that demand for Delaware in-corporation is affected by (1) possible agency costs in the lawyer-clientrelationship; (2) the quality of legal rules in the home-state; (3) afirm's need for regulatory influence in its home state; and (4) firm-specific factors (such as firm characteristics or planned transactions).Note, however, that I can make only cautious claims about what actu-ally causes a particular incorporation choice because firm domicile islikely endogenously determined with other firm characteristics such asownership structure, takeover defenses, planned transactions, and thelegal advisor. Below I try to account for the endogeneity of the domi-cile and legal advisor and to control for other firm-specific features.Section A of this Part describes possible theories of firm demand forDelaware law, and Section B presents regression results.

A. Theories of Firm Demand: What Firms Want

1. Agency Costs

Although keenly alert to agency costs that arise when firms hiremanagers, corporate law scholars are frequently silent about agencycosts that arise when firms hire lawyers. Instead, the state-competi-tion debate has generally assumed that managers get the law they pre-fer in a roughly Coasean world of low transaction costs and well-informed parties. In the classic Winter and Cary debate, managerssought (and got) either good or bad laws. However, managers do notchoose domicile themselves but instead hire others (lawyers) to advisethem.84 Once we allow that managers rely on agents, it is natural toask whether agents' self-interests might affect their advice.

One agency-cost hypothesis is that local lawyers encourage localincorporation because it is in their interest, while "national" law firms(i.e., those with clients in many states) encourage incorporation inDelaware. To see why this might be, imagine that Easterbrook wants

84 Underwriters may also provide input, and subsequent regressions control for them.Anecdotal evidence from intereviews with underwriters suggests, however, that they donot typically make this decision or take an interest unless the suggested decision is"unusual."


[Vol. 77:1559


to take his firm public. His current lawyer, Fischel, is influential inIllinois, knows Illinois law, and has a primarily local practice. Whatare Fischel's incentives when it comes to recommending a domicile forEasterbrook's firm? Other things equal, if Fischel recommends thatEasterbrook incorporate in Illinois, Easterbrook will need to hire Illi-nois lawyers in the future for advice on transactions and for litigationand Fischel will face less competition for Easterbrook's corporate-law-related business from outsiders who are not allowed to practice Illi-nois law (specifically, local incorporators would exclude Wall Street orlarge out-of-state firms that specialize in Delaware law). Illinois in-corporation also avoids the need for Easterbrook to hire (and Fischelto share fees with) Delaware lawyers in any litigated dispute. Thus,on the margin, because Illinois incorporation increases Easterbrook'sdemand for Fischel's services, Fischel may recommend local incorpo-ration, which may or may not be in the Easterbrook's interests.

Imagine another lawyer, Cary, a New Yorker with a large na-tional practice who knows Delaware but not Illinois law. Cary has theopposite incentives and may recommend that Easterbrook incorpo-rate in Delaware. First, having his nationwide clients in a commonjurisdiction (like Delaware) limits Cary's need to remain up to date onthe laws of many other states, lowers his cost of doing business, andutilizes his investment in learning Delaware law.85 By recommendinga Delaware incorporation, Cary also reduces the expected value ofknowing Illinois law and lawmakers, thus reducing the relevance ofIllinois law, which is Fischel's comparative advantage in competing forEasterbrook's future business. Of course, Cary may also be better atproviding advice regarding Delaware law and may recommend Dela-ware incorporation for that reason.

Thus, on the margin, it may be in the self-interest of local lawfirms to encourage local incorporation and national law firms to en-courage Delaware incorporation. The important thing is to note thepossibility that this may be a conflict of interest. Obviously, there arelimits to the extent that such agency costs can determine a firm's dom-icile. Professional norms may help reduce these agency costs by rein-forcing a lawyer's duty to seek her clients' welfare instead of her own.Competition between law firms will also reduce such agency costs,since lawyers who ignore their clients' interests will bear some cost.However, because clients have difficulty monitoring the quality of thelegal advice they receive, norms and competition are likely to mitigate

85 See Romano, supra note 13, at 275 (noting "the intuition that the interests of legalprofessionals have a significant impact on a Delaware domicile" and that outside counselservicing clients in many different states have incentive to encourage Delawareincorporation).



but not erase residual conflicts of interest (just as competition in thelabor market does not eliminate managerial agency costs).8 6

There are several important caveats to this: First, the choice oflawyer and domicile may be simultaneously determined. That is, Eas-terbrook may know that Delaware law is better for his firm and thatCary provides better advice on Delaware law. Thus, it is at least possi-ble that Easterbrook searches out Cary in order to obtain Delawarelaw for his firm. I try to account for this possibility using regressiontechniques described below.8 7 Nevertheless, there is other evidencethat suggests that lawyers determine domicile. Romano found thatreincorporation decisions were typically motivated by lawyers ratherthan managers. a8 Second, a large fraction of firms that do not incor-porate either at home or in Delaware incorporate in a nearby state(e.g., a Connecticut firm incorporating in New York).8 9 This is greaterthan the fraction that would occur if clients surveyed the national legallandscape by themselves and simply selected a lawyer to match.Rather, it is consistent with the idea that entrepreneurs know a locallawyer who practices in a nearby state, and therefore suggests incor-poration in that state. Moreover, anecdotal evidence suggests that cli-ents rely on legal advisors rather than independent preferences forlegal regimes. In conversations with lawyers, underwriters, and entre-preneurs, IPO participants did not suggest this as a credible alterna-tive. Thus, though this sort of lawyer selection is impossible to ruleout, I consider it unlikely.

Second, there may also be more benign interpretations of thesame behavior. If a client selects a local lawyer to advise on an IPO,the local lawyer may do a better job and provide better legal advice ifthe firm incorporates locally. The local lawyer's preference for locallaw may be the product of the lawyer's imperfect information ratherthan greed. Regression techniques will be unable to discern a lawyer'smotivation. Therefore, though the welfare effects are indeterminate,if Delaware law really did improve firm value between 1981 and 1996(perhaps by as much as 2% on average) 90 a local lawyer would haveneeded to add a great deal of value to make up for the mean reduc-tion in firm value from incorporating outside Delaware.

86 See William J. Carney, The Production of Corporate Law, 71 S. Cal. L. Rev. 715,721(1998).

87 See infra notes 110-13 and accompanying text.88 See Romano, supra note 13, at 274 tbl.13.89 See infra tbl.8 and accompanying text.

90 See Daines. supra note 7, at 533; Subramanian, supra note 13, at 1809.


[Vol. 77:1559


2. Quality of State Law

When a firm goes public, the quality of its state corporate lawsmay matter: If the state has high-quality corporate law, the firm maybe more likely to stay home. If its laws are of poor quality, the firmmay be more likely to incorporate in Delaware. Thus, good local cor-porate law may attract firms, whereas bad local corporate law mayrepel firms. The question of which laws attract or repel firms is at thecenter of the state-competition debate; examining which attract or re-pel firm incorporation will tell us what firms are looking for in corpo-rate law. This will in turn suggest something about the laws stateshave an incentive to produce: If takeover laws of a certain sort in-crease the number of firms incorporated in a jurisdiction, then statesmay have marginally more incentives to pass such takeover laws.91

The alternative hypothesis is that state law (outside of Delaware) istrivial and/or uniform and that variation in legal rules will not affectfirms' domicile decisions. 92

I examine four qualitative measures of a state's corporate laws.First, recent scholarship suggests that the value of corporate law rulesis partly a function of the number of firms adopting them.93 Just asthe value of a particular operating system or a fax machine is partly afunction of the number of users who adopt it, the value of a particulardomicile may increase with the number of firms incorporated there.This is known as a network effect. For instance, as the number offirms incorporated in a state increases, the amount of litigation in-creases, which may clarify the law, provide valuable training to judges,and increase the availability of cheap legal advice. In addition, themore firms that incorporate in a given state, the more likely the legis-lature may be to update the state's corporate code. In this view, theoverall value of a legal regime is the function of its inherent value (i.e.,the value of a product standing alone) and its network value (its value,given that others use the product).

If corporate law has such network benefits, then the value ofhome-state law will increase (and the likelihood of Delaware incorpo-

91 States, however, may have little financial interest in maximizing the number of firmsincorporated within their borders. See generally, Kahan & Kamar, supra note 14 (manu-script at 11-30). It is, of course, true that Delaware is trying to increase the number ofincorporations. Thus, if certain kinds of laws keep firms incorporated at home, this wouldaffect Delaware's incentives if not other states' incentives.

92 See, e.g., Black, supra note 10.93 Marcel Kahan & Michael Klausner, Path Dependence in Corporate Contracting: In-

creasing Returns, Herd Behavior and Cognitive Biases, 74 Wash. U. L.Q. 347, 350-58(1996); Klausner, supra note 83, at 843-44 ("[A]s the number of firms incorporated in astate increases, the value of its charter increases."); Romano, supra note 13, at 277-78 (ana-lyzing effects of increased incorporations).



ration will decrease) with the number of firms domiciled in each firm'shome state.94 If network effects are economically insignificant, thelikelihood of Delaware incorporation will be unaffected by the num-ber of firms incorporated in the home state.95

A second, related measure of the quality of a state's corporatelaw is its adoption of the Model Act.96 First, because roughly half ofthe states have adopted the Model Act, it may serve as a corporatelaw network. By adopting it, a state taps into a wide body of prece-dent and commentary, thus perhaps clarifying legal rules and reducingthe cost of legal advice relative to a stand-alone legal regime. TheModel Act also has, in some respects, different substantive rules thanmany states, and these substantive differences, rather than any net-work benefits, may also affect the value of the Act.97

As a third measure, I examine the degree to which the states'laws provide sufficient legal protection to allow the firm to sell to dis-persed shareholders. If a state's legal rules allow managers and con-trolling shareholders to exploit investors, investors will require pre-IPO shareholders to retain a larger fraction of the firm's shares.98 Ifmanagers begin to exploit investors, the price of the firm's shares willdrop and managers will bear the cost. Thus, absent adequate legalprotection, managers and entrepreneurs can credibly commit to taking

94 This is an imperfect proxy as it may also capture something about the inherent qual-ity of the law-as opposed to its network effects. In tests below, I try to distinguish be-tween the network and inherent quality aspects. See infra note 115 and accompanyingtext.

95 It may also be that incorporating in a jurisdiction with many other firms is valuablebecause it provides the firm with some assurance that there will be enough firms to lobbythe legislature in case the state's code needs to be updated. See Romano, supra note 13, at277 ("The more corporations there are, the more the state relies on their business and willrespond with desired legislation."). This argument is similar to the network argument-numbers are valuable-but it describes the source of the advantages differently.

96 Model Bus. Corp. Act, at xxvii (1998/1999 Supp.).97 Romano, supra note 13, at 278, argues that firms value a state's record in quickly

updating its corporate code. Thus, states that respond quickly to legal innovations mayfare better in the race against Delaware. I therefore examine the effect of a state's respon-siveness (based on Romano's 1985 data) on the likelihood that a firm located in that stateincorporates in Delaware, and expect responsiveness and Delaware incorporation to benegatively related. In regressions below, I find that the likelihood of local incorporation isgenerally increasing in state responsiveness, but the effect is dependent on the form of theregression, perhaps because of the different time periods in our studies. Romano's respon-siveness index is based on corporate law developments though the 1970s, while I am prima-rily examining IPOs in the 1980s and 1990s.

98 See Armando Gomes, Going Public Without Governance: Managerial ReputationEffects, 55 J. Fin. 615, 616 (2000). A similar result is obtained by Leland and Pyle, whoshow that managers will retain a large block of stock as a signal of the firm's quality. SeeHayne E. Leland & David H. Pyle, Informational Asymmetries, Financial Structure, andFinancial Intermediation, 32 J. Fin. 371, 384 (1977). However, such a signal is less valuableif legal protection reduces investors' uncertainty about investing in the firm.


[Vol. 77:1559


valuable actions by retaining a large block of shares at the IPO. Bycontrast, if a state's legal rules reduce agency costs, entrepreneurs willbe able to sell more stock to the public. Thus, other things beingequal, firms subject to better governance may have more widely dis-tributed share ownership. If Delaware law improves the value ofholdings of dispersed shareholders, other things being equal, IPOfirms that incorporate in Delaware will be more widely held. There-fore, in regressions below, I control for the percentage of the firm soldto the public in the IPO. Finally, I examine the impact of a state'sadoption of antitakeover statutes. Antitakeover laws have received agreat deal of attention from legal scholars who suggest that the adop-tion of these statutes may make a domicile more or less attractive.99

It is difficult to see, however, how these takeover statutes wouldaffect the domicile choices of IPO or reincorporating firms. First,most antitakeover statutes are default only and allow IPO firms tocostlessly opt into or out of important provisions. Given that firmsgoing public can easily add or subtract these simple provisions in theircharter without regard to whether the state has adopted such a stat-ute, it is difficult to understand how these statutes might affect a firm'sdomicile choice. 00 Second, most statutes examined are, as a substan-tive matter, now trivial given the development of the poison pill.Most antitakeover statutes provide protection that is no better than(and often less effective than) a poison pill, which all firms may freelyadopt on their own.' 0 ' Thus, many of these protections are likely triv-ial as a substantive matter, and are easily replicated or avoided as aformal matter, regardless of the state's statute. For these reasons, it ishard to see how most antitakeover statutes would be relevant to afirm going public.

Several states have adopted statutes that may matter, however,because they contain provisions that firms are unable to create con-tractually, such as statutes authorizing the adoption of a poison pill,constituency statutes, control-share acquisition statutes, or disgorge-

99 See Bebchuk & Cohen, supra note 33, at 4 ("[Ajntitakeover protections are corre-lated with success in the incorporation market."); Subramanian, supra note 13, at 1826-64(arguing that managers respond to antitakeover laws in choosing domicile).

100 Bebchuk and Cohen suggest that the legislature's adoption of (or failure to adopt) a

takeover statute sends a meaningful signal about the state's commitment to provide an-titakeover protection in the future. See Bebchuk & Cohen, supra note 33, at 19. The valueof such a signal is likely to degrade over time, however, and it seems unlikely that theadoption or non-adoption of particular statutes in the 1980s provides a meaningful signalof the state's corporate law more than a decade later. In most states, it seems likely that alarge employer could lobby and win entrenching legislation if it were helpful.

101 See James C. Coates IV, Takeover Defenses in the Shadow of the Pill: A Critique ofthe Scientific Evidence, 79 Tex. L. Rev. 271 (2000); Daines, supra note 70, at 4-5.



ment provisions.102 In regressions below, I control for states thatadopt these provisions.

3. Home-Court Advantage

Firms may incorporate at home in order to give managers influ-ence over the firm's corporate law rules. If this ability to influencelegislators and policymakers is useful, it is most likely used in take-overs-where the stakes for the firm are high and where firms havehad success in lobbying state legislatures. If so, then home-state incor-poration should be more likely when IPO firms are more likely toadopt takeover defenses. The likelihood of home-state incorporationwould then increase along with the threat of takeover (as measured bythe amount of takeovers in the firm's industry) and decrease alongwith research and development (R&D) intensity.10 3 Firms may alsoincorporate at home to secure other regulatory benefits. If so, firmssubject to important state regulation should be more likely to incorpo-rate in-state. In regressions below I control for public utilities, whichare subject to intense state regulation.10 4

4. Firm Transactions and Conflicts

The advantages of Delaware (or home-state) law may depend ona firm's business strategy. Romano argues that Delaware law will bevaluable to firms that plan acquisitions. 10 5 In regressions below, Itherefore control for the number of acquisitions each sample firmmade following its IPO.106

102 See generally Matheson & Olson, supra note 34.103 Takeover defenses in IPOs are generally less frequent in industries where research

and development expenses are large (relative to firm size). See Daines & Klausner, supranote 1, at 104-06.

104 See infra tbl.7. Initial results suggest that the number of employees is positively re-lated to the likelihood of home-state incorporation, consistent in a crude way with the ideathat firms incorporate at home when they have local political clout.

105 Romano, supra note 13, at 250. Delaware law is said to reduce acquisition costs fortwo reasons. First, a Delaware acquiror firm needs only the approval of a simple majorityof its shareholders, rather than a supermajority. Second, Delaware's thicker precedentallows managers to plan acquisitions to avoid liability.

106 See infra tbl.7. The interpretation of this result is complicated by the fact that Dela-ware law may, in turn, affect the number of acquisitions a firm makes. To limit this con-cern, I count only the bids made during the first three years following the IPO. During thistimeframe, shareholdings are relatively more concentrated and the shareholders, ratherthan formal corporate law, likely dictate the outcome and incidence of major transactions.Moreover, IPO firms presumably evaluate merger opportunities at the time they go public.Limiting the transactions to this timeframe limits, but does not eliminate, endogeniety con-cerns. Because the law of the target's domicile rather than the acquiror's governs most ofthe key questions in a transaction, I also examine the number of times each sample firmbecame the target of another bid in the three years following the IPO in unreported regres-sions without changing the basic results. The receipt of a quick bid is insignificantly related


[Vol. 77:1559


Similarly, some firms may have greater need for fiduciary dutyrules to constrain possible manager or shareholder opportunism. Forinstance, some firms go public with a controlling shareholder withwhom the firm will continue to do business. Spin-offs fit this descrip-tion. A spinoff occurs when a parent company (e.g., Ford) sells stockin a subsidiary (e.g., Hertz) to the public. In many cases, the parentcontrols the newly public subsidiary through its ownership and repre-sentation on SpinCo's board of directors. Often, the parent firm willcontinue to have dealings with SpinCo (Ford sells cars to Hertz),which raises the possibility that the parent will use these transactionsto extract above-market prices from the subsidiary thereby harmingshareholders of the sub. In such situations, conflict and litigation maybe more likely. Thus Delaware's body of precedent, specialized chan-cery court, and tradition of interpreting fiduciary duty rules may berelatively more valuable. I therefore control for sample firms that arespin-offs.

Finally, Delaware may also attract firms with valuable growth op-portunities. An earlier study found that Delaware firms are worthmore than firms incorporated elsewhere between 1981 and 1996. Oneexplanation of this finding is that Delaware attracts high-growthfirms.10 7 I attempt to control for this possibility in the regressions be-low by including the firm's sales growth in the three years followingthe IPO. In addition, on the grounds that firm quality and under-writer quality are likely to be correlated, I include two indirect mea-sures of quality: the presence of venture capital investment, and thequality or reputation of the firm's underwriter, following a commonlyused index.'08

B. Analysis and Regression Results

Table 6 below summarizes some of the basic differences betweenfirms that incorporate in Delaware and those that incorporate else-where. On average, Delaware firms are larger, more levered, andmore likely to make acquisitions and receive takeover bids in the firstfew years following the IPO. These findings are consistent with the

to Delaware incorporation, suggesting that the greater frequency of bids for Delawarefirms may not just be the product of an IPO selection bias.

107 Daines, supra note 7, at 551-53. This explanation is possible, but the theory behindsuch a selection bias is unclear.

103 Carter et al., supra note 48. Underwriters serve as reputational intermediaries andvouch for the quality of the firm's growth prospects, management team, and disclosures, aswell as screen potential clients.



view that large firms that are likely to undergo complicated transac-tions find Delaware law valuable.109

Firms are more likely to stay "at home" if their home state hasadopted the Model Act. Although Delaware's share has also in-creased over time-from roughly 40% to more than 70% over thesample period-its gain has largely come at the expense of states thatdo not adopt the Model Act, which lost roughly half their share toDelaware over the sample period (from 62% in 1980-1984 to 33% in

TABLE 6DESCRIPTIVE STATISTICS

1980-1989 1990-1999

Delaware DelawareFirms Other Firms Firms Other Firms

Financial characteristics (in $millions)Assets 304.27** 62.93 456.63** 142.15

(32.38) (16.28) (50.61) (34.18)Annual sales 150.03** 50.63 224.35** 102.54

(33.52) (17.36) (41.07) (30.33)Market cap 126.87** 56.56 192.21** 112.07

(47.04) (23.45) (90.39) (57.70)Debt/assets ratio .16** .12 .14** .11

(.08) (.05) (.03) (.02)Insider ownership after IPO - - 41.66** 45.78

(42.8) (47.8)TransactionsAverage number of acquisitions 2.39** 1.79 2.19** 1.97

(1) (1) (1) (1)Number of bids received .90* .81 .42** .31

(1) (1) (0) (0)Days firm remained independent 2,149** 2,484 983 1,044

(1,933) (2,484) 879 (928)Law firmNumber of states in which the firm 6.15** 4.9 7.9** 5.4has represented an IPO firm (5) (2) (6) (3)OtherUnderwriter quality 7.92** 7.39 8.12 7.41

(8.75) (7.88) (8.75) (8.5)% of IPOs with VC investment 49.74%** 56.86 40.75 31.74Big (5) auditing firmNumber of firms incorporated 154.45** 142.73 237.10 204.97in home state (55) (54) (82) (79)Number of firms headquartered 63.98** 58.64 88.40 77.33in home state (127) (107) (164) (140)

Differences in means are significant at the 1% (**) and 5% (*) levels, respectively.Numbers in parenthesis represent medians.

109 Another explanation is that Delaware law generates bids and acquisitions. To limitthis concern, I examine only bids in the first three years of the IPO. During this time, thefirm is usually still controlled by a small number of shareholders who own a controllingblock. Corporate law will not have any impact on the acquisition decisions of such firmsbecause the firm is not vulnerable (by reason of its ownership structure) to a hostile take-over attempt. Moreover, the control shareholders will determine whether or not to ap-prove any transaction.


[Vol. 77:1559


late the 1990s). By contrast, Model Act states retained roughly 50%of all their firms throughout the time period. This is consistent withthe idea that state law matters and that firms find the Model Actvaluable.

I use regression analysis to examine firms' incorporation deci-sions, controlling for a wide variety of factors identified above. Giventhe bimodal structure of the market for corporate law, I model theincorporation decision as a discrete choice whether or not to incorpo-rate in Delaware (using a logit model). The model is complicatedsomewhat by the fact that the firm's domicile and legal advisor willlikely be determined jointly or simultaneously, that is, we cannot un-derstand a firm's domicile decision in isolation but need to considerthe joint decision about domicile and lawyers.110

In order to clarify the relationship (and the potentially causal roleof lawyers), I employ a technique known as two-stage or simultane-ous-equations model. I first estimate a firm's demand for a nationallaw firm, controlling for variables that might be correlated with afirm's need to retain a large firm with a national practice-some ofwhich should be unrelated to the firm's demand for a particular domi-cile.11' I then use the resulting estimate of the degree to which a firmwill retain a "national" law firm as an independent variable in a sec-

110 The domicile decision is of course endogenous to a wide variety of firm governanceprovisions, but for tractability I focus on the relationship between lawyer and domicile. Itis also one of the strongest associations I find in initial and unreported regressions. Whilenational lawyers may advise clients to incorporate in Delaware, one alternative interpreta-tion is that some firms have plans or characteristics that make it desirable to have nationallawyers and, for similar reasons, also find it desirable to be in Delaware. For instance, afirm that anticipated merger and acquisition activity may need both Delaware law and alarge firm to advise it, because merger and acquisition work requires specialized legal tal-ent possessed by large firms who also attract clients from across the country.

111 I control for a wide variety of factors in this first stage. I attempt to control for afirm's ease of access to a large national law firm (e.g., entrepreneurs in New Jersey may bemore likely to end up with a large New York law firm than entrepreneurs in South Da-kota). Because firms in populous states may have easier access or more contacts withnational lawyers, I also control for the number of firms headquartered in the IPO firm'sstate. Firms that plan complicated transactions after going public may retain a large lawfirm, so I also control for acquisitions that occur in the first three years following the IPO.Delaware law is relatively unimportant to acquirors and likely did not affect theprobability of such transactions that could have been anticipated at the time of the IPO. Inaddition, I control for whether or not the firm received a bid in the three years followingthe IPO and a variable that measures the amount of takeover activity in the firm's industryduring the three years surrounding the IPO. I also control for firm size (using log of salesor market value) because large firms may complete "bigger" deals and require bigger lawfirms. I control for venture capital investment, the firm's initial sales growth following theIPO, and the reputation of the underwriter, on the grounds that large firms or firms withgrowth opportunities may also require the advice of experienced transactional (national)lawyers. Finally, I include separate dummy variables for the firm's industry, the year of theIPO, and each of the fifty states in order to capture any state-specific effects. Although



TABLE 7THE DEMAND FOR DELAWARE INCORPORATION

This table reports the results of discrete choice logit regressions estimating whether a firm incor-porates in Delaware. A positive coefficient estimate indicates the variable is associated with ahigher probability of Delaware incorporation.

2 stageLogit Logit Logit Logit 2 stage 2 stage 2 stage stepwise

Legal advisorNational law firm

State lawModel Act

Number of firms incorp.in state

Poison pill statuteControl share statuteConstituency statuteDisgorgement statuteFair price statuteBus. combination statuteFirm and transaction vari% of shares sold in IPO

Spinoff

Regulated utility

Post-IPO acquisitions

Industry merger activity

Firm size (market value)

Underwriter ranking

Sales growth (first 3years)

R&D

Headquarter statedummies

Industry dummiesYear dummiescorrect- 2 log likelihoodN

Significant at the 1% (***

0.07*** 0.06*** 0.06*** 0.06*** 0.34*** 0.34*** 0.21* 0.30***(0.05)

-0.66*** -0.61*** -0.00 -0.23* 0.02**(0.13)

-0.19*** -0.20*** -0.15 -.23*** -0.09(0.05)

0.30** 0.51** 0.52**0.66** 0.64*

-0.16 -0.22 -0.14-0.35 -0.16 -0.02

0.31 -0.240.22 -0.27

ables0.58** 0.62** 0.62** 0.73*** 0.73*** 0.63***

(0.28)0.61*** 0.61*** 0.66*** 0.74*** 0.74*** 0.82***

(0.17)1.71*** 1.70*** 1.86*** 1.54*** 1.82*** 2.53***

(0.53)0.10 0.11 0.08 0.01 0.00

(0.09)-0.00* -0.00* -0.00* -0.01* -0.01"*

0.000.00 0.00 0.00 -0.00 0.00

(0.00)0.18*** 0.18*** 0.18*** 0.01 0.07

(0.03)

0.00* 0.00** 0.00** 0.00 0.00(0.00)

0.66* 0.66 0.90** 0.66* 0.97**(0.40)

No No No Yes 1' stage 1' stage YesYes Yes Yes Yes Yes Yes YesYes Yes Yes Yes Yes Yes Yes

66.2 67.1 67.95,548 3,360 3,169 3,401 2,579 3,303 3,4224,612 2,825 2,825 2,825 2,914 2,914 2,905 2,914

*), 5% (**) and 10%(*) level.


[Vol. 77:1559


ond regression that estimates the firm's state of incorporation. In ad-dition to the variables described above, I control for the possibilitythat Delaware's relative attractiveness varies by industry (with con-trols for the firm's two-digit SIC code) and over time (with controlsfor the year of the IPO). I control for firm size by including the firm'smarket value, but similar results are obtained when I instead use thelog of the firm's sales. I also include headquarter-state fixed effects,including separate dummy variables for a state's headquarters. Thisallows lawyer and domicile to be jointly and simultaneously deter-mined, avoiding potential endogeneity concerns. 112 Regressions in-clude fewer observations due to missing variables.

Results are summarized in Table 7 above. The first four columnsin the table present results of one-stage logit regression estimates. Be-cause the results of these regressions are very similar to the two-stagemodel, I primarily discuss the latter. I first control only for law firm,industry, and year of IEP0. For robustness checks, I then successivelyadd state-law variables, firm-specific variables, and finally takeover-statute variables and state-fixed effects.

In each model, the estimated effect of a national law firm is posi-tive, which means that the likelihood of Delaware incorporation in-creases when the law firm has a more national practice. The effect ofthe law firm is generally highly significant, both statistically1 3 (p-valuetypically = 0.0001-which means that there is less than one chance in10,000 that the relationship is by chance) and economically (on aver-age, an increase of five states in which the law firm has clients is asso-ciated with an approximately 50% greater likelihood of Delawareincorporation). Moreover, the addition of the law-firm variable doesmore to improve the fit of the model than any of the variables control-ling for a state's law, more than the controls for a firm's characteris-tics, and more than all other variables combined. Thus local lawyersadvise firms to incorporate locally, while national lawyers advise firmsto incorporate in Delaware. This is consistent with the idea that law-yers act in their self-interest in providing this advice. An alternativeinterpretation is that lawyers provide their best advice but do not fullyunderstand the advantages of Delaware (or home-state) law.114

this technique deals with endogeneity concerns, it is impossible to conclusively rule themout.

112 If the domicile is jointly determined with the identity of the lawyer, estimates of firmdomicile that ignore this endogeneity and the effect of the lawyer would be incorrectlyspecified and potentially misleading.

113 In all of the regressions I run, it is statistically the most significant of all relationships(p-values = 0.0001).

114 The two-stage regression attempts to deal with the possibility that certain firms (forwhom Delaware is optimal) simply seek out national law firms.



There is some mixed evidence consistent with the theory that cor-porate law has network qualities. In columns B and C home states aremore attractive when large numbers of public firms are already incor-porated there. The relationship is highly significant (p-value =0.0001). Thus, other things equal, it appears that the value of a state'slaw increases with the log of the number of firms incorporatedthere. 115 Moreover, firms are more likely to incorporate in homestates that adopt the Model Act, which functions as something of anetwork of precedent and commentary. An alternative explanation isthat firms value the Model Act's substantive rules, and I cannot dis-criminate econometrically between these two interpretations.' 1 6

This relationship, however, is sensitive to the form of the regres-sion and the variables I include. I obtain inconsistent results when Iinstead use the absolute number of incorporated firms rather than thelog, suggesting that the effect of the number of firms may be non-linear. Controls for the number of firms incorporated in-state becomeinsignificant in the presence of separate dummy variables for individ-ual home states, and in OLS regressions estimating the proportion oflocal firms retained by each of the fifty states (in unreportedregressions).

I find weak support for the idea that firms incorporate in-state inorder to secure some regulatory advantage. Firms incorporate in-statewhen the threat of takeover is higher, which suggests that they mayvalue the ability to lobby for takeover protection. Industry mergerand aquisition activity is negatively related to Delaware incorporation

115 An alternative explanation of this result is that states that adopt good substantiverules attract more incorporations. This is possible and cannot be excluded. However,there are several reasons to think we observe network-rather than inherent-benefits.First, I try to control for important substantive differences between states (such as theadoption of the Model Act or various takeover statutes), though there are likely unobserv-able differences. Second, in unreported regressions, I also include controls for a state'ssuccess in retaining locally headquartered firms (its retention rate)-as well as for the ab-solute number of firms headquartered in-state. If, in the presence of these controls, thenumber of firms incorporated became insignificant, but the retention rate was significant,this would suggest that the quality of the legal rules rather than the number of firms incor-porated (and thus the network effects) was critical. However, the number of firms incor-porated remains significant even in the presence of these controls, so I do not reject thenetworks explanation.

116 If the substantive differences between the Model Act and states' corporate law re-gimes are relatively small, the better interpretation is that this suggests network effects. Inunreported regressions, I find that there is some evidence that firms going public value astate's responsiveness in updating its corporate code, as Romano's measure of responsive-ness is negatively related to the likelihood of Delaware incorporation (p-value = 0.001). Inalternative specifications, the relationship is sometimes insignificant. This is not inconsis-tent with Romano, supra note 13, as I use her measure of responsiveness calculatedthrough 1984, while it is possible that state responsiveness changes over the course of mysample period.


1596 [Vol. 77:1559


(the effect is significant at the 10% level). However, the effect ofR&D intensity is insignificant and regulated public utilities are actu-ally less likely to incorporate at home. I do not control for firm-leveltakeover defenses, however, so I cannot rule out the possibility thatthese firm-level defenses vary between Delaware and non-Delawarefirms. However, in a separate sample of IPO firms drawn with MikeKlausner, I found no difference in the mean firm-level defensesadopted by Delaware or non-Delaware firms.

I find little support for the claim that antitakeover statutes affectthe domicile choices of IPO firms. Poison pill or control share statutesmake it more likely that firms will leave a state and incorporate inDelaware. However, this effect is not robust and is theoretically ques-tionable given the ease with which firms can opt into or out of theseprovisions. In unreported regressions I examine alternative measuresof a state's takeover laws and find that coefficient estimates are unsta-ble and sensitive to the measure and choice of control variables. Itherefore present results with and without these controls.

I also attempt to control for biases introduced by the fact thatfirms are not equally distributed across states, but are concentrated ina few large states (California, Illinois, New York, etc). Other studiesof state takeover statutes, such as Bechuck and Cohen's, 117 have ex-amined simply the choices of individual firms. However, because thismethod overweights the impact of large states, regression results saymore about large states than about a system of relations between thefifty states. To correct for this bias, I estimate each state's success inretaining local firms, thus equally weighting each state. In these unre-ported regressions, takeover statutes are again insignificant or in-crease the likelihood of Delaware incorporation. This suggests thattakeover statutes are not important in attracting incorporations andthat takeover statutes do not provide evidence that states are "racingto the bottom.""18

Spin-offs are more likely to incorporate in Delaware, consistentwith the idea that firms incorporate in Delaware when they are likelyto engage in transactions that may require rulings on managers' fiduci-ary duties. The fraction of the firm's shares owned by public share-holders is positively associated with the likelihood of Delawareincorporation. This is consistent with the idea that Delaware lawreduces agency cost in public firms and protects investors. If investors

117 See Bebchuck & Cohen, supra note 34.118 This is inconsistent with Bebchuk & Cohen, supra note 33, who examine the impact

of takeover statutes on a different sample of firms. I further examine takeover statutes inanother paper, but initial results indicate that my results are unchanged even using alterna-tive models and measures of takeover statutes.



have fewer protections in other states, they would require a greaterdegree of concentration in order to provide entrepreneurs with theincentive to reduce agency costs. Another interpretation is that ifDelaware law improves the value of the firm's shares, a firm's foun-ders have greater incentives to sell shares to the public. Because thefirm's ownership structure is also endogenous to the firm's incorpora-tion, the interpretation of this relationship is uncertain.

To confirm these results and check their robustness, I also esti-mate a stepwise logit regression. Stepwise regressions add or subtractpotentially important independent variables according to the varia-ble's effect on the model's predictive strength. This resulting modelincludes the variables with the strongest effect on Delaware incorpo-ration but has the limitation that it allows the data (rather than the-ory) to specify the ultimate model. I use it here to check on therobustness of the associations. Results are presented in Column H ofTable 7. Consistent with prior results, the variables of theoretical in-terest (lawyers, networks/precedent, conflict transactions) are statisti-cally significant and retain the expected sign.

C. Other Evidence

One way of examining how firms choose a domicile is to examinefirms that do not fit the general pattern-i.e., firms that incorporateoutside of their home state but not in Delaware. In particular, thesefirms can highlight the motivation for domicile choice and why firmsleave the home state-whether these firms leave because they aresearching for particular legal rules. I therefore separately examine allIPO firms between 1995 and 1997 that incorporated neither in Dela-ware nor in their home state. I then collected a copy of the prospectusfrom each firm's IPO. This produced a subsample of eighty-one firms.

Interestingly, many of these firms have a physical connection totheir domicile: The domicile is an adjoining state (e.g., some Connect-icut or New Jersey firms incorporated in New York), the state of thefirm's parent company, or the state where the firm began major oper-ations. Many firms began in the state, or have some original connec-tion with the incorporating state (i.e., it being the location of the firm'sparent company or founder) and some retain major operations in thedomicile state.119 See Table 8 below. Given their tangible, physical tieto the domicile state at the firm's inception, these firms are best seen

119 In a spinoff, the parent firm sells shares in the subsidiary to the public. Of these

spunoff firms, 38% of all firms not incorporated locally are incorporated in the home stateof their parent firm.


1598 [Vol. 77:1559


as quasi-home-state incorporations, rather than firms that searchedfor and found a particular set of technical corporate law rules.

The fact that firms are often incorporated in a nearby state is con-sistent with the idea that lawyers influence incorporation decisions.An entrepreneur in Connecticut is more likely to know (and hire) aNew York lawyer than a California lawyer. Therefore, if lawyers wereimportant in domicile decisions, we would expect to see firms incorpo-rating in nearby states. This is also consistent with the idea that law-yers matter and that lawyers suggest a local domicile at start-up.

I find additional evidence that lawyers affect incorporation deci-sions when I examine two different subsamples of firms. First, I ex-amine only the subset of firms that incorporate locally. If locallawyers too frequently recommend local incorporation, firms repre-sented by local lawyers will be less responsive to qualities that makeDelaware attractive. That is, if the likelihood of Delaware incorpora-tion generally increases with underwriter reputation, given local coun-sel, underwriters will need to be of higher reputation before locallawyers recommend Delaware incorporation. Conversely, if a giventrait makes local law attractive, it will take more of that trait beforenational lawyers recommend local incorporation. I find some evi-dence of this: Firms represented by local lawyers come from smaller,less responsive states, that are more likely to have adopted the ModelAct, sell a greater fraction of the firm's stock, and be taken public byan underwriter with a better reputation (variables, as explained above,are otherwise associated with Delaware incorporation). This is consis-tent with the idea that local lawyers recommend local incorporation tofirms that would otherwise find Delaware attractive. Finally, to ac-count for the possibility that sophisticated managers of relatively large

TABLE 8WHY DO FIRMS INCORPORATE IN A THIRD STATE?

% of firms w/% of Total Firms known connection

Type of firm in 3' state to 3rd state

Quasi-instate incorporationsFormer headquarters, major manufacturing

facilities, or regulated sales within the state ofincorporation 43.2% 50.0%

Firm was spunoff from parent corporationheadquartered or incorporated in the state ofincorporation 9.9 11.4

Firm was founded in the state of incorporation 6.2 7.1Adjacent state 18.5 24.4Director links to the state of incorporation 2.5 2.9Other 6.1 7.1Unknown 13.6 N/A



firms may select national lawyers and independently prefer Delawarelaw, while other firms incorporate locally, I examine only the subset offirms with prior venture capital investment. Results are essentiallyunchanged in unreported regressions.

CONCLUSION

The debate over state corporate law has focused almost exclu-sively on theoretical reasons for believing that Delaware is racing tothe top or the bottom and on evaluating Delaware's prominence.There is, however, surprisingly little evidence about how firms actu-ally make incorporation choices at the IPO stage, how the market forcorporate charters is organized, or how other states or regulatoryschemes fare.

In this Article, I examine the incorporation choices made at theIPO, the starting line of any corporate law race and before ownershipis dispersed and agency costs arise. I document the death of Tieboutcompetition. Though firms are free to incorporate in any of the 50states, they rarely do: Firms incorporate either at home or in Dela-ware. No other state has been able to attract out-of-state incorpora-tions and no state attracts more than 1/2 of 1 percent of the assets offirms seeking legal rules out of state.

Thus, a primary finding of this Article is that the dominant meta-phor for state corporate law - a national race to the top or the bottomor a market with 50 producers-is incorrect and potentially misleading.Instead, the market looks more like a series of 49 paired-duopolies (orrun-offs) between Delaware and other states.120 There is simply nomeaningful actual competition for national firms outside of Delaware;its lead in the national market is overwhelming. This suggests that thecrucial question at this point is not simply whether states are racing tothe top or the bottom-but what kind of race or market this is and whatfactors or barriers to entry sustain the equilibrium I observe.

One implication of these facts is that there is a home bias orhome-court advantage in incorporation decisions. I examine two po-tential explanations for this home bias: a local firm's ability to lobbyfor corporate law reform (particularly antitakeover statutes) and therole and possible agency costs of lawyers. I find evidence more consis-tent with the idea that lawyers matter. Lawyer identity appears toexplain more of the variation in firm decisions than any other factor,including the substance of the state's legal rules. Local lawyers appearto suggest local incorporation, and lawyers with multi-state practicesuggest Delaware incorporation. One potentially troubling aspect of

120 Any national market is potential only.


[Vol. 77:1559

December 2002] INCORPORATION CHOICES OF 1PO FIRMS

this is that the advice matches their own self-interest, although I can-not exclude other interpretations. 121 However, given evidence thatDelaware firms were more valuable and more likely to receive take-over bids in this sample period, it seems reasonable to suspect that thisaffects shareholder welfare.

With these facts on the table, future work will examine the impli-cations of these findings for corporate law. A few areas of inquiryappear profitable, however. First, one must ask how much of a racethere is. Delaware's lead is enormous and growing (its share of IPOsrose from about 40% to 80% in the sample period). Other states'complete failure to attract out-of-state firms (including Nevada's fail-ure) suggests the difficulty of competing with Delaware and may dis-courage other states from attempting to compete in the long run.Moreover, if each state's market is effectively limited to local firms,states (especially small states) have relatively little incentive to "com-pete" for charter fees or to invest in innovations. Thus, although Del-aware's small size and industrial base is an advantage and allows it tobecome dependent on franchise taxes and limits the political influenceof takeover targets,122 for other states, being small is a disadvantage.In the alternative, the absence of a meaningful number of firms incor-porating outside their own boundaries anywhere but Delaware, maybe taken as evidence consistent with the idea that other states are notcompeting for incorporations. Thus, Delaware may face few actualcompetitors and enjoy considerable slack in the race to the top/bot-tom and market power in the market for corporate charters. In thewords of Ralph Winter, it may be a "leisurely walk" rather than avigorous national competition.'2 3 This would be relatively good newsif competition's critics are generally correct, as Delaware would facefewer incentives to quickly race to the bottom. However, this wouldbe relatively bad news if competition's supporters are generally cor-rect, as Delaware and other states would have fewer incentives to raceto the top.

This data also raises questions about the direction of any racebetween states. If lawyers affect the demand for corporate law (as

121 Though I do not evaluate the welfare effects of this conflict in detail, evidence sug-gests that unless the local lawyer was dramatically less expensive, on average firms mayhave been better off in Delaware. This is, of course, only an average and there is no reasonto think that Delaware incorporation was optimal for every firm.

122 See Daines, supra note 7, at 540.123 Ralph K. Winter, The "Race for the Top" Revisited: A Comment on Eisenberg, 89

Colum. L. Rev. 1526, 1529 (1989).



they do its supply),124 any race is not driven only by manager-share-holder conflict and, at least in this respect, is neither to the top nor thebottom. That is, even if states are trying to increase incorporations,they would not simply craft laws to suit managers' preferences, butwould rationally cater to lawyers (whose interests may or may not co-incide with the firms' interests).

One key question is whether this market structure (and homebias) affects shareholder welfare. Given the valuation and transactionconsequences of Delaware incorporation, it seems that the choice be-tween Delaware and home state incorporation matters. However, it isless clear that much turns on the choice between a firm's home stateand, say, Oregon. If non-Delaware states produce largely similar cor-porate law, the fact that firms ignore third states is unsurprising andthe home bias is untroubling. If corporate law outside of Delaware islargely fungible, even a small in-state benefit would be sufficient toproduce a home state bias. Differences may or may not be meaning-ful. Given that lawyers appear to affect the choice between homestate and Delaware where there is apparently variation, it is not im-plausible that potentially meaningful variation is overlooked oro\avoided by the parties or lawyers. It is not obvious that serial (orpair-wise) competition generates a different or less efficient laws thannational head-to-head competition with 50 states. Even given thismarket structure, Delaware would still compete in New York for NewYork firms, and in California for California firms. Thus, Delawaremay adopt similar legal rules regardless of whether it competes in onelegal market or 49 distinct legal markets. One potential implication ofthese findings, however, is that Delaware may focus on potential com-petition from populous states where most of their potential customersare located. Moreover, the threat of potential competition constrainsDelaware.125 Thus, the market structure suggests that Delaware en-joys meaningful market power, but does not without more compel theconclusion that the resulting corporate law rules are inefficient.

I find some mixed evidence that network benefits may explainboth individual firm decisions and Delaware's large and sustained

124 See Carney, supra note 86, at 720-21 (1998) (arguing that corporate lawyers can in-fluence production of corporate laws); Macey & Miller, supra note 30, at 473 (arguing thatDelaware lawyers shape Delaware law with their own interests in mind).

125 Even absent a nationwide competitor with substantial share, there are three likelysources of competition: (a) home states, which by default would become more attractive;(b) the federal government (which has sometimes intervened in state corporate law mat-ters, such as the Williams Act, 15 U.S.C. § 78(a)-(mm), and the recent Sarbanes-Oxley Actof 2002, Pub. L. No. 107-214, 116 Stat. 745 (codified in scattered sections of 11, 15, 18. 28,and 29 U.S.C.); or (c) states adopting the Model Act.


[Vol. 77:1559


lead.126 The welfare implications of this one unclear. If network ef-fects are important, there is reason to be less optimistic about the con-straining effect of potential competition from other states. It wouldbe extremely difficult for a state to gain share by developing particularstatutory schemes for the simple reason that legal rules cannot becopyrighted. If a state adopted a legal rule valuable enough to enticeexisting firms, Delaware could simply imitate the rule, making it op-tional. Thus, firms would have no incentive to leave Delaware for theother state's substantive law. This could discourage states from inno-vating in corporate law reform.127 As a result, a state interested ingaining market share would only be able to compete over networkbenefits-which is highly unlikely to be successful given Delaware'senormous installed base. 128 Small states would likely be ineffectivecompetitors, given the small number of firms located there (eventhough a limited revenue stream might otherwise lead to a depen-dence on charter revenues and thus make the state a good competi-tor). 129 Thus, network effects and the ability to copy corporate lawrules may help explain Delaware's persistent lead. If so, then one can-not conclude that the observed outcome is the optimal one. On theother hand, even if individual states cannot provide network benefitsto compete with Delaware, they can effectively band together andprovide a competing network by adopting the Model Act. Moreover,large states may attract sufficient numbers of local firms that they be-gin to create network benefits of their own. Thus, even recognizingnetwork effects does not mean that Delaware faces no competition.

Finally, whatever its problems, firms' preference for Delawaredoes not appear to be pathological or the result of managerial agencycosts. During this sample period, Delaware attracted the most firmsand appears to have offered valuable legal rules and to be relatively

126 Network effects may also explain why lawyers do not more frequently learn the lawof a third state. See supra text accompanying notes 79-80.

127 In fact, this is one way to interpret the evidence that Delaware typically does notinnovate in corporate law reform, but rather is an early adopter of other states' innovationsSee Romano, supra note 13, at 260.

128 States might also compete on price, perhaps even offering a subsidy (negative price)to incorporating firms. States have not done this to date. Another potential method ofcompetition would be to provide a specialized chancery court.

129 See Romano, supra note 13, at 278 ([O]nly if a substantial number of firms could agree to move to a state otherthan Delaware in a coordinated fashion, could that state begin to compete withDelaware. It is the existence of a large pool of resident corporations that...allows [Delaware] to develop a legal system that promotes the sale of chartersby reducing the costs of doing business to firms.).


1603


unlikely to entrench incumbent managers. 130 Delaware IPO firmswere also more widely held, consistent with theories that Delawareimproves governance and reduces agency costs in public firms. Fi-nally, firms with reputable underwriters, venture capital or leveraged-buyout investment were also more likely to end up in Delaware, sug-gesting that it is not just chosen by the ill-advised. Given this, it is atleast as important to examine the actions of other states and firms thatdo not incorporate in Delaware. This article is a first step in examin-ing these choices.


130 Daines, supra note 7, at 541-42 (noting that Delaware's political economy limits man-agers' ability to obtain entrenching laws).

[Vol. 77:1559


APPENDIX

TABLE AlSTATE SUCCESS OR FAILURE IN ATTRACTING

OUT-OF-STATE PUBLIC FIRMS

This table examines a state's success in attracting incorporations among a sample of publiclytraded firms.

Out-of-state firms

incorporated in state %

DE 1155 83.57

NY 51 3.69

NV 29 2.10

NJ 22 1.59

MD 16 1.16

VA 12 0.87

PA 10 0.72

MA 10 0.72OH 9 0.65

TX 6 0.43

N 1382 100.00

Out-of-state firms

incorporated in state %

DE 1963 87.79

NY 51 2.28

NV 50 2.24

NJ 18 0.81

MD 17 0.76

PA 14 0.63MN 11 0.49

MA 11 0.49VA 10 0.45

IN 10 0.45

N 2236 100.00



TABLE A2STATE SUCCESS OR FAILURE IN RETAINING IPO FIRMS

This table examines a state's success in attracting incorporations of locally headquartered firmsamong a sample of [ I IPO firms between 1980-2000.

1980-1989 1990-1999

Local firmsStates with best incorporated Firms HQ'dretention rates in-state % in-state

DE 100 4

WA 67 24

MI 67 15

OR 64 11

NV 63 8

MN 63 19

AR 60 5

WI 59 17

IN 57 7

NC 53 15

GA 50 26

OK 50 4

NM 50 2

CA 49 260

OH 45 31

MA 45 66

States with worstretention rates

AK 0 1

HI 0 1

ID 0 2

RI 0 2

WV 0 3

DC 0 3

CT 4 24

AZ 10 10

LA 11 9

IL 12 41

MO 13 16

UT 17 6

NH 20 5

VA 21 24

FL 23 22

Local firmsincorporated Firms HQ'din-state % in-state

DE 100 5

OR 69 36

NV 67 21

MN 65 72

LA 53 17

WA 52 77

WV 50 2

SD 50 4

ND 50 2

MT 50 4

GA 48 100

FL 47 143

TN 47 58

MI 47 38

IN 46 28

OH 45 78

AK 0 3

AR 0 7

HI 0 3

ID 0 5

NM 0 3

VI 0 1

WY 0 1

CT 4 74

DC 7 15

NH 8 13

AZ 12 41

IL 12 118

NY 12 208

CO 13 80

MO 15 33


[Vol. 77:1559


TABLE A3STATE SUCCESS OR FAILURE IN RETAINING PUBLIC FIRMS

This table examines a state's success in retaining incorporations among a sample of [ ] IPO firmsin 1985 and 1995.

Local firmsStates with best incorporated Firms HQ'dretention rates in-state % in-stateVT 100 2

NV 93 14

DE 86 7

OR 83 18

IN 78 41

WI 76 37

HI 75 4

MN 73 90

ME 67 3

OH 60 119

SC 60 10

PA 58 111

GA 58 65

10 58 12

MA 57 117

States with worstretention rates

MT 0 1NWV 0 2

DC 0 4

IL 7 122

NH 9 11

ID 13 8

AR 15 13AL 18 17

AZ 19 16

CT 20 85

CA 25 325

OK 26 23

KY 27 11

MO 28 43

TX 28 209

Local firmsincorporated Firms HQ'din-state % in-state

DE 100 11

NV 81 32

MN 74 136

IN 71 49

OR 68 38

SD 67 6

ME 67 3

WI 66 65

OH 55 139

WA 55 51

MI 55 78

IA 50 18

HI 50 4

AK 50 2

PA 49 156

MT 0 4

DC 0 6

NH 0 18

IL 9 164

ID 11 9

CT 13 103

AL 14 29

AZ 15 39

AR 16 19

VT 20 5

WV 20 5

TX 23 338

MD 24 49

NM 25 4

CA 25 551


1607


TABLE A4THE MARKET FOR IPO CHARTERS: 1990-2000

A B C D E F

Number of Out-of-state Market share Number offirms Market firms of all firms that firms

incorporated Share incorporated incorporate out headquartered RetentionState instate % in-state of state in state rate %

AK 0 0 0 0 3 0

AL 4 .12 0 0 16 .25

AR 0 0 0 0 7 0

AZ 5 .15 0 0 41 .12

CA 223 6.66 9 .37 899 .24

CO 13 .39 3 .12 80 .13

CT 4 .12 1 .04 74 .04

DC 2 .06 1 .04 15 .07

DE 2298 68.66 2293 95.3 5 1.00

FL 71 2.12 4 .17 143 .47

GA 53 1.58 5 .21 100 .48

HI 0 0 0 0 3 0

IA 5 .15 2 .08 12 .25

ID 0 0 0 0 5 0

IL 18 .54 4 .17 118 .12

IN 16 .48 3 .12 21 .46

KS 5 .15 0 0 14 .36

KY 2 .06 0 0 12 .17

LA 10 .30 1 .04 17 .53

MA 64 1.91 0 0 247 .26

MD 32 .96 14 .58 64 .28

ME 2 .06 0 0 5 .40

MI 20 .60 2 .08 38 .47

MN 55 1.64 8 .33 72 .65

MO 5 .15 0 0 33 .15

MS 3 .09 0 0 13 .23

MT 2 .06 0 0 4 .50

NC 24 .72 1 .04 52 .44

ND 1 .03 0 0 2 .50

NE 4 .12 1 .04 12 .25

NH 1 .03 0 0 13 .08

NJ 26 .78 3 .12 99 .23

NM 0 0 0 0 3 0

NV 38 1.14 24 1.00 21 .67


[Vol. 77:1559

December 2002] INCORPORATION CHOICES OF IPO FIRMS 1609

Number of Out-of-state Market share Number offirms Market firms of all firms that firms

incorporated Share incorporated incorporate out headquartered RetentionState instate % in-state of state in state rate %

NY 32 .96 8 .33 208 .12

OH 36 1.08 1 .04 78 .45

OK 7 .21 0 0 24 .29OR 27 .81 2 .08 36 .69

PA 47 1.4 5 .21 119 .35

RI 2 .06 0 0 10 .20

SC 3 .09 0 0 15 .20

SD 2 .06 0 0 4 .50

TN 29 .87 2 .08 58 .47

TX 65 1.94 0 0 306 .21

UT 11 .33 1 .04 23 .43VA 21 .63 4 .17 80 .21

VT 1 .03 0 0 6 .17

WA 42 1.25 2 .08 77 .53

WI 13 .39 0 0 29 .45

NWV 1 .03 0 0 2 .5

WY 0 0 0 0 1 0N 3347 100 2406 100 3347 -



TABLE A5THE MARKET FOR CORPORATE CHARTERS IN PUBLIC FIRMS: 1995

A B C D E F

Number of Out-of-state Market share Number offirms firms of firms firms

incorporated Market incorporated incorporating headquartered RetentionState in-state Share in-state out of state in-state rate %

AK 1 .03 0 0 2 .50

AL 4 .11 0 0 29 .14

AR 3 .09 0 0 19 .16

AZ 6 .17 0 0 39 .15

CA 142 4.03 3 0.13 551 .25

CO 31 .88 6 0.27 82 .30

CT 15 .43 2 0.09 103 .13

DC 1 .03 1 0.04 6 0

DE 1974 55.98 1963 87.79 11 1.00

FL 78 2.21 10 0.45 147 .46

GA 52 1.47 6 0.27 103 .45

HI 3 0.09 1 0.04 4 .50

IA 10 .28 1 .04 18 .50

ID 2 .06 1 .04 9 .11

IL 18 0.51 3 .13 164 .09

IN 45 1.28 10 .45 49 .71

KS 11 0.31 2 .09 22 .41

KY 7 0.20 2 .09 14 .36

LA 11 0.31 3 .13 22 .36

MA 100 2.84 11 .49 193 .46

MD 29 0.82 17 .76 49 .24

ME 2 0.06 0 0 3 .67

MI 45 1.28 2 .09 78 .55

MN 111 3.15 11 .49 136 .74

MO 23 0.65 4 .18 70 .27

MS 4 0.11 0 0 14 .29

MT 0 0.00 0 0 4 0

NC 31 0.88 3 .13 58 .48

ND 31 0 0 0 0 .31

NE 5 0.14 1 .04 13 .06

NH 1 0.03 0 0 18 .26

NJ 57 1.62 18 .81 149 .25

NM 1 0.03 0 0 4 .81

NV 76 2.16 50 2.24 32 .34


[Vol. 77:1559

December 2002] INCORPORATION CHOICES OF IPO FIRMS 1611

Number of Out-of-state Market share Number offirms firms of firms firms

incorporated Market incorporated incorporating headquartered RetentionState in-state Share in-state out of state in-state rate %

NY 147 4.17 51 2.28 281 .55OH 85 2.41 8 .36 139 .35

OK 13 0.37 2 .09 31 .68

OR 28 0.79 2 .09 38 .49

PA 90 2.55 14 .63 156 .36

RI 4 0.11 0 0 11 .47SC 9 0.26 0 0 19 .67

SD 4 0.11 0 0 6 .43

TN 26 0.74 2 .09 56 .23

TX 82 2.33 5 .22 338 .46

UT 16 0.45 3 .13 28 .39

VA 42 1.19 10 .45 82 .20

VT 1 0.03 0 0 5 .55

WA 30 0.85 2 .09 51 .66

WI 47 1.33 4 .18 65 .20

WV 1 0.03 0 0 5 0WY 2 0.06 2 .09 0 0

N 3526 100 2236 100 3526 -


incorporation choices of ipo firms, the - nyu law … · the incorporation choices of ipo firms ......

Documents